Oxidative stress/toxicity (OT) in humans is a function of partial pressure of oxygen and time of exposure (1) and is thought to be caused by damage from oxygen free radicals (2, 3). It has been shown to occur in nearly all tissues and organs (4). The most sensitive target organs are the lungs and central nervous system (1) (CNS: brain and spinal cord) and the manifestations are usually separated into acute toxicity from continuous exposure and chronic toxicity from repetitive exposures. Acute CNS OT from continuous exposure to oxygen was first described in animals in 1878 with hyperbaric exposures and is known as the Paul Bert Effect (5). Acute pulmonary OT from continuous exposure to oxygen was first described in 1899 with normobaric and hyperbaric exposures and is known as the Lorrain Smith Effect (6). Chronic or cumulative pulmonary OT from repetitive exposure at hyperbaric pressures has been described and led to the development of the unit pulmonary toxic dose (UPTD) metric in divers (7–9). Chronic or cumulative CNS OT from repetitive hyperbaric exposures has been implied and replicated in animals (10–12), but dismissed/attributed to other causes in humans (13–16). It has only been described from continuous exposures of hours to days and consisted of paresthesias and numbness in fingers and toes, headache, dizziness, nausea, and reduction in aerobic capacity (17).

The research and clinical experience on CNS OT under hyperbaric conditions has been conducted and reported at ≥2.0 ATA oxygen on neurologically normal subjects (18–20), or emergency indication patients (13–15). These studies report the extreme manifestation of acute CNS OT, grand mal seizures. Due to the low incidence of seizures at 2.0 ATA and the traditionally greater clinical levels of 2.36–2.4 ATA (13–15) CNS OT under hyperbaric conditions was considered non-existent or impossible to elicit below 2.0 ATA (21–26) or more recently 1.9 ATA (27). Even when obvious CNS symptoms developed from continuous exposure at 2.0 ATA “…the cause of the symptoms is not known” (22) or the symptoms were characterized as a “somatic” constitutional form of OT (28), omitting the descriptor “chronic” to avoid confusion with chronic pulmonary OT (29). The rectangular hyperbola pressure- continuous time exposure graph of acute CNS OT supports this belief (30) and has been reinforced by a lack of identification and reporting of cumulative CNS OT with repetitive HBOT under 1.9–2.0 ATA. This may be due to the reporting of CNS OT from the common clinical use of HBOT at pressures of 2.0 ATA or greater (31–39), but more likely is due to the lack of application of HBOT to chronic neurological conditions prior to Neubauer (40).

Based on Neubauer’s work (40), the absence of reported CNS OT, and belief of non-existence of CNS OT below 1.9–2.0 ATA (21–27), Harch, Van Meter, and Gottlieb investigated the treatment of chronic neurological disorders with 40 treatment blocks of HBOT at 1.5 ATA/90 min daily from 1994 to 1999 (41). Cases with unanticipated untoward symptoms emerged in a group of carbon monoxide- poisoned patients with neuro-cognitive residual symptoms treated 6 months post carbon monoxide exposure (Reported at the Undersea and Hyperbaric Medical Society Gulf Coast Chapter Meeting 4/1995, New Orleans). Additional cases from multiple sources at exposures >1.5 ATA oxygen, prompted a cautionary statement to a worldwide parent support group for disabled children, the MUMS Network, in 1998 (reprinted 2022) (42) and a presentation of cumulative cases at the 2nd International Symposium on Hyperbaric Oxygenation and the Brain Injured Child in 1999 (43). In the subsequent 24 years the increasing neurological off-label use of HBOT by untrained medical and non-medical personnel, home use of hyperbaric therapy, continued contact of the primary author by patients with negative outcomes from HBOT for neurological disorders, and reports of permanent injury and death indicated a need for more widespread understanding of safe dosing limits of HBOT for chronic neurological disorders. This retrospective study addresses this need. Herein we report acute CNS OT within 5 HBOTs and chronic CNS OT with repetitive HBOT exposures, both at less than 2.0 ATA oxygen.

Medical Records Archive and Search: The primary author maintained archives of all medical records of his hyperbaric patients from the time of the first application of hyperbaric oxygen therapy to divers with chronic decompression illness (September 29, 1989) and of all communications since 9/1994 from outside sources (physicians, hyperbaric medical facilities, families, and all individuals who sought consultation regarding hyperbaric oxygen therapy, particularly for complications of hyperbaric treatment) by phone, email, letter, or from a subsequent formal medical history and physical exam. The entire communication archive was searched from 9/1994 through Institutional Review Board approval of January 20, 2023 and a partial search of the medical records, mostly from the author’s memory and separately archived cases with complications, was conducted from September 29, 1989 to January 20, 2023.

Type of Records Selected: Records were selected that featured signs and symptoms (SS) of oxygen toxicity described by Donald (19, 20) that occurred during treatment with HBOT for chronic neurological conditions. These SS were change in level of consciousness (loss of consciousness, drowsiness, sleepiness, amnesia, confusion, loss of judgment), neuromuscular signs (clonic spasms of the legs, “blubbering of the lips,” muscle fasciculations, clumsiness, fibrillation of lips, lip twitching, twitching of cheek and nose or any muscle), salivation, behavioral SS (changes of behavior, fidgeting, disinterest), mood changes (depression, euphoria, apprehension, acute terror), autonomic SS (facial perspiration, diaphoresis, facial pallor, bradycardia, palpitations, sensation of arterial pulsation throughout the body, syncope); respiratory SS: panting, grunting, hiccoughs, inspiratory predominance, spasmotic respiration, choking sensation, epigastric tensions, epigastric aura, constriction or same in precordium; alteration of senses: paresthesias, visual symptoms (flashes of light, haloes, loss of acuity, lateral movement of images, decrease of intensity, constriction of visual field, hallucinations, micropsia), acoustic symptoms (music, bell ringing, knocking), hallucinations (unpleasant olfactory or gustatory sensations); malaise, spasmotic vomiting, vertigo, and SS in the acronym VENTIDS (visual, ear, nausea, twitching, irritability, dizziness, seizure). Synonyms were allowed, e.g., lethargy, lassitude, fatigue, and brady-kinesia for drowsiness or sleepiness.

Data Extraction, Analysis, and Statistics: Data from enrolled cases were extracted and a clinical vignette composed for each case. Data included age, sex, diagnosis, time from injury to HBOT, the HBOT schedule/dose, number of HBOTs at time of OT symptom/sign onset, the SS of OT, and cumulative oxygen dose at the onset of SS. Cumulative dose was calculated in atmosphere-hours (AHs) with the following equation:

Cases were identified as the authors’ or from outside sources and were sorted by their historical chronological occurrence. The first cases were observed at 1.5 ATA during the 1994 clinical trial (vide supra). They were segregated from cases treated at > or < 1.5 ATA oxygen. Chamber type was designated as hardshell (steel and acrylic, typically, with a pressure limit of 3.0 ATA or greater) or softshell (synthetic soft material composition and pressure limits of 1.3 ATA). Functional brain imaging was included when available. SPECT scanner specifications, radiopharmaceutical, acquisition, and processing are described in Supplementary File S1 (44, 45). This study was approved as LSU IRB #4574. GraphPad¹ was used to perform unpaired t tests on group comparisons. Means were calculated using Calculator.net.²

Eighty-two cases were enrolled. Twenty-two cases were excluded due to equivocal signs and symptoms, critical missing data, or attribution to alternative explanations of SS, leaving 60 cases for the final cohort. These comprised 27 cases from a previous IRB-approved study and 33 cases since 7/2001. The study was closed at 60 cases, arbitrarily judged by us to be sufficiently representative of chronic CNS OT.

Abbreviated histories of all cases are in Supplementary File S2. Case reports from outside sources used the mother, family member, or immediate caregiver’s quotation of signs and symptoms. During the review acute CNS OT cases were identified from outside clinics. Five HBOTs were chosen as the cutoff for acuity (32–35). Included in the chronic cases were patients who demonstrated initial improvement in neurological and cognitive function with HBOT then reversed these improvements with continued HBOT, re-expressing their neuro-cognitive deficits, and finally regained their HBOT- induced improvements after discontinuance of HBOT. This pattern strongly suggested the cumulative oxidative stress with HBOT and recovery from same on discontinuance of HBOT described by Bean and Siegfried (46). The pattern was also indicative of typical drug overdosing and recovery upon discontinuance of the drug. As a result, this pattern was identified as a CNS OT manifestation.

Data from the Supplementary File S2 case reports are presented in Supplementary Tables S1–S5 and in Supplementary Table S1 (acute oxygen toxicity), Supplementary Table S2 (chronic CNS oxygen toxicity at 1.5 ATA oxygen), Supplementary Table S3 (chronic CNS oxygen toxicity at >1.5 ATA oxygen), Supplementary Table S4 (chronic CNS oxygen toxicity at <1.5 ATA oxygen), and Supplementary Table S5 (single case of withdrawal syndrome). Data from Supplementary Tables S1-S4 are condensed in Table 1.

Supplementary Table S1 contains 7 cases of acute CNS OT, most of which (5 cases) occurred at ≤2.0 ATA. They were 2–95 y.o., 6 males, 1 female, with 7 different neurological diagnoses (two with seizure disorders). Five of the cases were in the chronic disease phase (greater than 6 months from time of injury, insult, infection, diagnosis). The single mortality was a 95 y.o. male with 3,000 previous HBOTs for a stroke who was treated a few hours after an acute stroke, experienced CNS OT on the second treatment, a grand mal seizure on the 3rd HBOT (1.1 ATA oxygen), 15 subsequent grand mal seizures in the next 24 h, and died after the 15th seizure.

Supplementary Table S2 lists 31 cases of chronic CNS OT after >5 HBOTs and at 1.5 ATA oxygen that include the first two unrecognized cases treated by the authors in 1990 and 1992 (47). The patients were 18 months to 60 y.o., 19 M, 12 F, 4 weeks to 10 years post injury/insult (29 ≥ 6 months since injury/insult), with 13 different neurological diagnoses. There were two severe complications from outside sources occurring in home hardshell chambers (one death and one institutionalization for severe behavioral deterioration). One case’s OT (#16) was serendipitously captured on SPECT brain blood flow imaging (Figure 1) in an IRB-approved study of HBOT in chronic severe TBI. There were nine cases with second episodes of CNS OT and six cases with three or more episodes.

Supplementary Table S3 contains 12 chronic CNS OT cases occurring at >1.5 ATA oxygen pressure. There were 9 males, 3 females, ages 2–58 y.o., 8d to 12 years post injury/insult, with 9 different diagnoses. There was one death from an outside source in a home hardshell chamber. The OT in six of 12 cases occurred after an escalation of oxygen pressure or on intensive twice/day or 6–7d/week schedules. Seven of the 12 cases occurred at 1.75 ATA, one at 1.55 ATA at 6,900 feet altitude, two at 2.0 ATA, one at 2.4 ATA, and one at 2.8 ATA. Second episodes of OT were documented in 5 cases and three or more episodes of CNS OT in one case.

Supplementary Table S4 contains nine chronic CNS OT cases occurring at <1.5 ATA oxygen pressure. There were 4 males, 5 females, ages 2.5–83 y.o., 2.5 months to 11 years after injury/insult with seven different neurological primary diagnoses. Four cases of OT occurred in portable chambers, two of whom occurred after prolonged chamber treatments (2–3 h each) or a prolonged course of treatment (135 HBOTs), and a fifth with post-surgical normobaric oxygen after long-term HBOT in a portable chamber. One was a subacutely drowned child who could not tolerate 1.15 ATA/45 min of compressed air. Overall, 4/9 had at least one episode while receiving compressed air at ≤1.5 ATA. Second episodes of CNS OT were documented in five cases and three or more episodes occurred in 3 patients.

Supplementary Table S5 lists a single case of apparent drug withdrawal that occurred after 130 HBOTs, the last 70 of which were consecutive, 7d/week, twice/day treatments. In the subsequent 2 weeks she experienced regression of her gains followed by a marked worsening of her pre-HBOT dystonia, involuntary total body movements, and teeth-grinding.

The critical data from Supplementary Tables S1-S4 are condensed in Table 1. The acute and chronic cases occurred over similar wide age ranges. Acute cases experienced CNS OT at <5 AHs while the chronic cases occurred regardless of pressure over a wide range of cumulative oxygen exposures with mean of 103–116 AHs. Second and third episodes of chronic CNS OT occurred with lesser additional cumulative AHs, markedly so in the cases at <1.5 ATA oxygen (approximately at 1/5th the AHs of the first episode).

Two comparisons were performed on the chronic CNS OT cases: (1) First episode in children (107 ± 94 AHs; case #49 omitted from calculation per above) vs. adults (117 ± 106): found to be non-significant (p = 0.72), and (2) Delay to treatment, based on start of continuous treatment that resulted in CNS OT. Subacute, ≤ 3 months (Cases #27, 48, 50, 57, 58: 21.1 ± 8.8 AHs), were compared to chronic cases, > 3 months [all other cases in Supplementary Tables S1-S4, except case #49 (46 cases)]: 123 ± 102 AHs, p = 0.035.

Using well-identified SS of acute CNS OT, 52 cases of chronic CNS OT from outside and internal sources are presented in this study. For comparison we also gathered 7 cases of acute CNS OT from ≤5 HBOTs. These 59 cases have demonstrated that: (1) The historical rectangular hyperbola of pressure/time continuous O2 exposure for CNS OT in normal people does not apply to patients with acute or chronic neurological conditions, (2) Acute CNS OT can occur at <2.0 ATA oxygen, (3) Chronic CNS OT exists and can occur at non-traditional lower pressures (<2.0 ATA oxygen and < 100% FiO2); (4) CNS OT can occur in chronic neurological conditions at very low doses, such as compressed air at 1.3 ATA, < 100% FiO₂ at 1.3 ATA, and even 1.15 ATA of compressed air in both portable and hardshell chambers; (5) Both acute and chronic CNS OT at <2.0 ATA occurred at 103–116 AHs, although there was wide variability consistent with the historical finding of acute CNS OT at higher pressures; (6) There was no difference in sensitivity between adults and children, but subacute intervention resulted in earlier CNS OT than delayed intervention; and (7) CNS OT was reversible if recognized early, but resulted in permanent morbidity and mortality if treatment persisted.

All cases were separated initially by pressure dose of HBOT at first occurrence of OT (cases at 1.5, > 1.5, and < 1.5 ATA oxygen and with less than 100% FiO₂ oxygen) based on their historical presentation in our initial clinical practice and then an IRB-guided study on HBOT in chronic neurological conditions using 1.5 ATA oxygen. We found that this historical separation was arbitrary. While there was a wide range of oxygen sensitivities at a given oxygen pressure consistent with Donald’s findings (19, 20), chronic CNS OT occurred at a nearly equal average cumulative oxygen dose (116, 103, and 114 AHs), regardless of oxygen pressure. Many of these cases (#‘s: 8,9,11,12,13,14,15,17,18,20,21,22,23,25,27,34,37,39,40,41,43,44,45,46,48,50,52,53,55,59) occurred after prolonged courses of treatment and/or from intensive twice/day or 6–7d/week schedules in monoplace chambers that commonly were not equipped or did not use air breaks, a well-proven method for avoiding oxygen toxicity (48–51). Twenty-one percent occurred after an increase in oxygen or pressure dose (case #‘s: 22,27,39,40,41,42,46,51,52,53,57). This was even manifest at very low doses such as pressurized air where CNS OT was experienced after increases in chamber pressurization time to 2–3 h. Regardless of the dose at which CNS OT occurred, the CNS OT was mitigated in some cases (Case #‘s: 17,19,22,46,53,55,56) by a decrease in oxygen exposure (decrease in intensity of schedule or reduction in dose) similar, again, to the well-proven effects of air breaks (48–51).

Our study also found that repetitive episodes of CNS OT occurred at lesser doses of oxygen consistent with animal studies which have shown sensitization to CNS OT after an initial episode (52). For the cases at 1.5 and > 1.5 ATA oxygen the second episode of CNS OT occurred at similar amounts of additional AHs, 81 and 67, respectively, and for those at 1.5 ATA the third or greater episode of CNS OT after 25 AHs. At >1.5 ATA there was only a single case of 3 or more episodes of CNS OT and so did not afford a comparison to the other oxygen pressures. At less than 1.5 ATA oxygen the second episode was experienced after 22 additional AHs, and the third or greater episode after 5.4 additional AHs. These lower numbers suggested an idiosyncratic hypersensitivity, reflecting Donald’s findings of wide idiosyncratic variability in oxygen sensitivity (19, 20).

Our study identified one case (case #60) of an habituation effect to HBOT where shortly after cessation of a prolonged intensive course of HBOT there was a notable deterioration in the patient’s condition to a level worse than the pre-HBOT level. This phenomenon was characteristic of a withdrawal syndrome seen with habit-forming drugs (53). Unfortunately, there was no long-term follow-up information on this case.

If CNS OT was recognized early and accommodated no significant lasting clinical injury was apparent. This comports with the widespread clinical impression of no harm from even the most severe form of CNS OT, seizures (1, 20, 54). However, Bert (5) believed that a toxic chemical substance was responsible for the OT seizures and Bean and Siegfried (46) conceived that the toxic substance of “transient or minute quantity” might still induce tissue injury (46). This transient toxic substance of minute quantity is now known to be reactive oxygen species (2, 55). ROS/oxidative stress have been shown to occur with single HBOT exposures at both 1.5 and 2.4 ATA (56). ROS and their by-products in OT seizures in animals have demonstrated permanent effects, including lipid peroxidation (3), apoptosis (57), cognitive injury (58), motor deficits (46), and ischemic lesions (11). Similar to Bean and Siegfried’s description (46), if oxidative stress continued in the setting of SS of CNS OT lasting injury occurred. In our series this included two deaths and in a third case who required institutionalization due to severe behavioral deterioration.

Bean and Siegfried (46) methodically investigated both acute and chronic CNS OT in animals, using repetitive (3–4x/day), multi-day 5.42 ATA oxygen exposures. Their findings included: (1) OT reactions during decompression and post-decompression could be described by acute and chronic phases, (2) Acute toxicity occurred dominantly during decompression or post-decompression and was idiosyncratic with a wide variation in individual susceptibility, (3) The intensity and duration of acute reactions showed wide variation and recovery was rarely instantaneous or immediate, (4) The susceptibility to seizures was increased by successive exposures and the time to occurrence of premonitory symptoms decreased with repeat exposures, (5) Successive exposures increased the severity and prolonged the duration of the acute post-decompression reactions such that lengthening the surface interval or stopping the exposures for a day or more was necessary for recovery, (6) Seizures began most commonly during decompression, (7) Individual sensitivity to the acute and chronic manifestations was unpredictable and empiric, (8) Acute decompression effects eventually merged with chronic effects which developed gradually and upon which repeat acute reactions were layered, (9) The outstanding feature of the chronic phase was motor manifestations that were permanent, (10) Augmentation of chronic effects could be induced by increasing the duration of the oxygen exposure or decreasing the surface interval, (11) In less susceptible animals the surface interval was important: increasing the frequency from 3 to 4x/day hastened the onset and decreased the total number of exposures necessary to bring about chronic alterations, (12) A few animals that were resistant to seizures became lethargic and anorexic, (13) The reactions to O2 were generally reversed by lowering the oxygen pressure or returning to normal atmospheric pressure.

Nearly all of the above findings of acute and chronic effects of OT are exhibited by the patients described in our study’s clinical vignettes. The character and development of these patients’ adverse SS are described by finding #‘s 1, 2, 3, 4, 5, 7, 8, 10,11, 12, and 13. The primary differences between our cases and Bean and Siegfried’s (46) animal cases are: (1) the initial manifestations were often more subtle, likely due to the lower doses of HBOT used, but accumulated with repetitive exposures, similar to what has been described in large reviews of patients developing oxygen toxicity seizures during courses of HBOT for standard indications at higher doses (31–36), (2) seizures occurred both during the treatment and in-between treatments, and (3) the outstanding features of the chronic phase were not only motor, but behavioral, emotional, affective, and other. While not tested by Bean and Siegfried (46) we also found no difference in development of chronic CNS OT between adults and children.

In addition, all of our study’s cases had neurologic comorbidity, implying an increased sensitivity to CNS oxidative stress with neurologic pathology. This sensitivity was more apparent in the patients with subacute neurologic injury where OT occurred at only 20% of the first episode cumulative AHs of patients with >3 months post injury. It suggests an augmentative effect of ROS on recently injured/inflamed neural tissue. The sensitization effect of neurologic comorbidity on CNS OT is consistent with the major review by Heyboer et al. (35), however, our cases occurred at a greater average cumulative oxygen exposure, 111 AHs, than the cumulative oxygen exposure (61 AHs) for the most severe form of oxidative stress, CNS OT seizures, in six large series of patients with mostly chronic HBOT wound indications treated at the typical 2.4 ATA/90 min (31–36). This is likely due to the higher pressures used in these series. These pressures and durations of exposure are near the inflection point in the rectangular hyperbola (30) of pressure/time continuous exposures for increased manifestation of CNS OT. Oxidative stress is proportional to pressure and time of exposure with mitigation by air breaks and surface intervals. With these greater pressures there is an increased level of oxidative stress with each treatment and cumulatively greater levels of oxidative stress require greater times to decay in both CNS (46) and pulmonary OT (1, 9). Given the similar air break (surface interval) of 22–23 h in a once/day treatment in most of our and the large series’ cases it is likely that there is retained injury from oxidative stress with these higher pressure treatments that has not been mitigated by the time of the next daily.

TREATMENT. This would be consistent with the kindling effect of repetitive exposures to hyperbaric oxygen (10) and the autocatalytic nature of oxidative stress (49) where these repetitive higher pressure exposures likely reach the auto-catalytic threshold sooner than patients exposed to the lower pressures of oxygen. This, again, reinforces Bean and Siegfried’s (46) proposal of transient effects of HBOT exposures and oxidative stress as well as permanent (and cumulative) effects which have been validated by others (10, 12).

Neurologic comorbidity in our subjects may also explain the most important finding of our study, that the rectangular hyperbola pressure/time CNS OT graph with its 2.0 ATA oxygen asymptote does not apply to patients with cerebral disorders receiving repetitive exposures to HBOT. The rectangular hyperbola misled us and the medical community into the false notion that CNS OT could not occur under 2.0 ATA oxygen. The central flaw in this notion was that the curve was generated from continuous hyperbaric oxygen exposures to normal people, mostly young male military members. We speculate that the human central nervous system in normal young males has the anti-oxidant capacity to equilibrate the oxidative stress of a continuous exposure to oxygen at less than 2.0 ATA, but not above. We further speculate that had seminal researchers in CNS OT (18–20, 22) exposed neurological patients to the same continuous exposures a typical hyperbola with both y and x axis asymptotes would have likely resulted. This is what our data describes with CNS OT occurring across a range of oxygen pressures from 2.0 to 0.24 ATA oxygen which approaches an x axis pressure asymptote. The only difference is that the autocatalytic threshold appears to be in the range of 103–116 AHs of cumulative oxygen exposure in chronic neurologic patients and far less with subacute/acute neurologic disease.

In one case CNS OT was captured on functional brain imaging that tracked the clinical manifestations and recovery (Figure 1). The acute manifestations were consistent with what has been identified with CNS OT in animals, increased brain blood flow prior to seizures (59). Despite the clinical improvement in the patient, the final imaging suggested negative residual effects with areas adjacent to the toxic areas showing significant reductions in blood flow below normal. Interpretation of this finding is difficult since the area of toxicity (high activity) would be expected to be the area showing OT injury (vide supra), not the adjacent areas unless there is an ischemic blood flow steal phenomenon from the adjacent areas to the high flow areas during the toxicity period.

An unexpected and derivative implication of the similar SS of CNS OT regardless of pressure and FiO₂ of oxygen is that increased pressure and hyperoxia are bioactive across the range of pressures from 0.24 to 2.0 ATA. This bioactivity, particularly for the patients who expressed SS of CNS OT at 1.3 ATA of compressed air, is additional proof that subjects receiving compressed air cannot be a control group in hyperbaric medicine clinical trials. The claim in HBOT cerebral palsy studies (60, 61) and mTBI PPCS studies (62–65) is that compressed air control groups are placebo groups. A placebo must be inert (66). If compressed air was inert it would be impossible for the subjects in our study to have experienced CNS OT.

Limitations of this study include its retrospective nature and inability to verify cases from outside sources with medical records or direct observation as in the authors’ cases. However, verification was discounted due to the context of the cases: parents of special needs children are keen observers, very familiar with their childrens’ neurological deficits, and the parents alarm occurred in the setting of expected beneficial outcomes. Deviations from these expectations were easily observed. Some data is missing and details of “neurological,” “cognitive,” “symptomatic,” and “physical” “improvements” are inexact in some instances. These omissions and inaccuracies were felt to be minor because the most important part of the vignettes were the patterns and negative SS exhibited by the patients which were similar. The patterns were either a neurological reversal after neurological improvement or development of SS of CNS OT identified for acute CNS OT that worsened as HBOT continued. Both the neurological reversal and CNS OT SS could not be attributed to any other cause and they recovered after discontinuation of HBOT, similar to acute CNS OT.

Lack of objective measures of oxidative stress/toxicity is a major limitation. Documentation with biomarkers would have strengthened the study. Only one instance of biomarker documentation was serendipitiously captured with SPECT brain imaging (Figure 1). Before OT biomarkers were available CNS OT was documented by observation of signs and reporting of symptoms. The list of these identified SS are very broad and amount to oxidative stress at almost any anatomical site in the CNS. The absence of biomarkers in this study is a distinct flaw, but one that is intrinsic to the historical habit of hyperbaric CNS OT documentation. Given the non- prospective nature of this study documentation with biomarkers would have required pre/post sampling which was impossible in a retrospective study. While many biomarkers of OT are available today (85–90), none are routinely used in clinical hyperbaric medicine practice, and the results are mixed in their use in clinical hyperbaric oxygen studies (91). Future clinical studies on hyperbaric CNS OT should employ biomarkers.

Other limitations include a possible contribution of increased barometric pressure and oxygen vasoconstrictive effects. A contribution of increased barometric pressure to what has always been assumed to be the effects of oxygen pressure alone are possible. Even slight increases in pressure are bioactive for nearly all living organisms (77), but any contribution to the observed SS in our cases is impossible to parse or assess in our study. Oxygen vasoconstrictive effects might explain some or all of the CNS OT SS in acute cases 4, 6, and 7 of Supplementary Table S1, but would seem unlikely in the 52 chronic cases where ischemic deterioration would be expected in the chamber on each HBOT. This is not what we observed or families reported.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

The studies involving humans were approved by the Louisiana State University School of Medicine Institutional Review Board. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because It was a retrospective chart review and due to the time over which the records were reviewed it would have been impossible to contact all of the subjects. The research would have been impossible to perform without the waiver.

PH: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing. SR: Data curation, Methodology, Project administration, Resources, Writing – original draft.