COVID-19 induced acute respiratory distress syndrome treated with Hyperbaric Oxygen: Interim safety report from a multicenter, randomised, open-label phase II clinical trial (COVID-19-HBO)

Abstract

Purpose: A few prospective trials and case series have suggested efficacy of hyperbaric oxygen therapy (HBOT) for treatment of severe COVID-19 but safety is a concern, especially for critically ill patients. We present the safety interim analysis of the randomised controlled trial COVID-19-HBO
Methods: Randomised controlled, open label, clinical trial in compliance with good clinical practice to explore the safety and efficacy of HBOT for severe COVID-19 in critically ill patients with moderate acute respiratory distress syndrome (ARDS). Between June 3 2020 and May 17 2021, 31 patients with Severe COVID-19 and moderate to severe ARDS; PaO2/FiO2 <26.7kPa (200mmHg) and at least 2 defined risk factors for ICU admission and/or mortality were enrolled in the trial and randomised 1:1 to Best practice or HBOT in addition to Best practice. Subjects allocated to HBOT received maximum 5 treatments 240 kPa, 80 minutes, during 7 days. Follow up was 30 days. Safety endpoints were analysed.
Results: Adverse events (AE) were common, hypoxia was most commonly reported, there was no statistically significant difference between the groups. Numerically, serious adverse events (SAE) and barotrauma were more frequent in the control group. Numerically differences were in favor of the HBOT in PFI, NEWS but statistically not significant at day 7, and no difference was observed for the total oxygen burden and Cumulative Pulmonary oxygen Toxicity Dose (CPTD).
Conclusions: HBOT appears safe as an intervention for critically ill patients with moderate to severe acute respiratory distress syndrome (ARDS) induced by COVID-19.
Trial registration: NCT04327505 (March 31, 2020) and EudraCT 2020-001349-37 (April 24, 2020)

Full text

Page 1/19
COVID-19 induced acute respiratory distress
syndrome treated with Hyperbaric Oxygen: Interim
safety report from a multicenter, randomised, open-
label phase II clinical trial (COVID-19-HBO)
Anders Kjellberg (  anders.kjellberg@ki.se )
Karolinska Institutet
Johan Douglas
Blekingesjukhuset
Adrian Hassler
Karolinska Institutet
Sarah Al-Ezerjawi
Karolinska University Hospital
Emil Boström
Karolinska University Hospital
Lina Abdel-Halim
Karolinska Institutet
Lovisa Liwenborg
Karolinska Institutet
Eric Hetting
Karolinska Institutet
Anna-Dora Jonasdottir-Njåstad
Blekingesjukhuset
Jan Kowalski
JK Statistics AB
Sergiu-Bogdan Catrina
Karolinska Institutet
Kenny A Rodriguez-Wallberg
Karolinska Institutet
Peter Lindholm
Karolinska Institutet
Article

Page 2/19
Keywords: COVID-19, respiratory distress syndrome, hyperbaric oxygen therapy, clinical trail
Posted Date: April 14th, 2022
DOI: https://doi.org/10.21203/rs.3.rs-1494203/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License. 
Read Full License

Page 3/19
Abstract
Purpose: A few prospective trials and case series have suggested ecacy of hyperbaric oxygen therapy
(HBOT) for treatment of severe COVID-19 but safety is a concern, especially for critically ill patients. We
present the safety interim analysis of the randomised controlled trial COVID-19-HBO
Methods: Randomised controlled, open label, clinical trial in compliance with good clinical practice to
explore the safety and ecacy of HBOT for severe COVID-19 in critically ill patients with moderate acute
respiratory distress syndrome (ARDS). Between June 3 2020 and May 17 2021, 31 patients with Severe
COVID-19 and moderate to severe ARDS; PaO2/FiO2 <26.7kPa (200mmHg) and at least 2 dened risk
factors for ICU admission and/or mortality were enrolled in the trial and randomised 1:1 to Best practice
or HBOT in addition to Best practice. Subjects allocated to HBOT received maximum 5 treatments 240
kPa, 80 minutes, during 7 days. Follow up was 30 days. Safety endpoints were analysed.
Results: Adverse events (AE) were common, hypoxia was most commonly reported, there was no
statistically signicant difference between the groups. Numerically, serious adverse events (SAE) and
barotrauma were more frequent in the control group. Numerically differences were in favor of the HBOT in
PFI, NEWS but statistically not signicant at day 7, and no difference was observed for the total oxygen
burden and Cumulative Pulmonary oxygen Toxicity Dose (CPTD).
Conclusions: HBOT appears safe as an intervention for critically ill patients with moderate to severe acute
respiratory distress syndrome (ARDS) induced by COVID-19.
Trial registration: NCT04327505 (March 31, 2020) and EudraCT 2020-001349-37 (April 24, 2020)
Take Home Message
The interim safety analysis of our randomised controlled trial, demonstrate a favorable safety prole in
acute respiratory distress syndrome induced by COVID-19. Safety concerns regarding barotrauma and
pulmonary oxygen toxicity (POT) were not supported and we speculate that HBOT may be useful to
reduce cumulative oxygen burden in ICU patients with ARDS.
Background
Severe COVID-19 often presents as an inammatory condition in the lungs (resembling organizing
pneumonia) and vascular endothelitis. The original aim of this study was to use the anti-inammatory
effects of Hyperbaric Oxygen Therapy (HBOT) to prevent intubation and save ICU-beds. In a wider
perspective we have investigated the use of hyperbaric oxygen in a severe pneumonia, near respiratory
failure. Despite decreased mortality in severe COVID-19, there is a need for additional safe and effective
treatments for patients developing acute respiratory distress syndrome (ARDS)(1). Our results may be
useful beyond COVID-19.

Page 4/19
HBOT consists of breathing 100% oxygen above normal atmospheric pressure raising inspired partial
pressure of oxygen (PO2) beyond 101.3 kPa as high as 280 kPa. This greatly increases oxygen transfer
and delivery through diffusion barriers. In addition to frank gas delivery the high PO2 has specic
biological effects; it reduces inammatory cytokines through several transcriptional factors, including
HIF-1(2, 3). By attenuating Nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB), possibly
through HIF-1, HBOT has the ability to restore inammatory homeostasis(4). HBOT is used in clinical
practice for several inammatory conditions such as radiation injury (5–7), ares of ulcerative colitis (8)
and diabetic foot ulcers (9). HBOT is associated with reduced mortality when used as adjuvant therapy in
severe bacterial infections such as necrotising soft tissue infections(10) and brain abscesses(11).
Although HBOT has virtually no relevant side effects in the regular patient populations, patients with
severe COVID-19 are often supported by high ow oxygen 24/7 creating a risk of pulmonary oxygen
toxicity (POT) that could potentially be accelerated by HBOT(12). Severe COVID-19 patients differ from
other ARDS patients by often presenting with “happy hypoxaemia” (hypoxaemia in the absence of
dyspnea, suggesting an adaption to low PO2), hence might be at greater risk of POT if traditional goals
for arterial partial pressure of oxygen (PaO2) is targeted(13). Another caveat is potential barotrauma from
gas expansion on decompression. Healthy lungs have traditionally been a requirement in diving and
unpressurized ight. In clinical practice, patients with severe COPD, pulmonary brosis, or cystic disease
are normally excluded from HBOT(14). There is an ongoing debate whether clinical equipoise exists for
this treatment in ICU patients with severe pulmonary disease(15).
HBOT for Severe COVID-19 was rst demonstrated in a case study from Wuhan, China, (16). Additional
reports including a randomised clinical trial have been published during the pandemic supporting
potential positive effects while not demonstrating any increase in adverse events (AE) for HBOT(17–20).
Several hypotheses with the common denominator “anti-inammatory effect” have been postulated(21,
22).
The present trial is ongoing and ecacy endpoints are not evaluated. The trial protocol has been
previously published(23). We hereby report on the safety prole to guide other researchers in trial design
and support clinicians that may consider HBOT for compassionate use in critical care patients with
pathological lung tissue.
Methods
Study design and overview
This phase II multicenter, randomised, controlled, parallel group, open-label clinical trial to evaluate safety
and ecacy of HBOT for severe COVID-19 is conducted at three sites: Blekingesjukhuset, Karlskrona and
Karolinska University Hospital, Stockholm, in Sweden and St Caritas University Hospital, Regensburg, in
Germany. The trial is investigator initiated and the sponsor is Karolinska Institutet, Solna, Sweden. The
trial complies with the International Conference on Harmonization of technical requirements for

Page 5/19
registration of pharmaceuticals for human use (ICH) Good Clinical Practice (ICH-GCP) and safety data
was reviewed by an independent data safety monitoring board (DSMB). The rst DSMB meeting was
conducted after 20 patients had completed the trial and the recommendation was to continue the trial
unchanged. The trial protocol has been previously published(23).
Between June 3 2020 and May 17 2021, 31 patients 18–90 years old with moderate to severe ARDS
induced by COVID-19; partial pressure of arterial oxygen/fraction of inspired oxygen (PaO2/FiO2) <
26.7kPa (200mmHg) and at least 2 dened risk factors for ICU admission and/or mortality were enrolled
in the trial at the Swedish sites. The CONSORT Flow Diagram for the safety analysis is shown in Fig.1.
Figure 1.
CONSORT ow diagram
Patients with severe COPD, signicant pulmonary brosis, or other contraindications for HBOT were
excluded(23). Follow up was 30 days.
Randomisation
Subjects were enrolled and randomised consecutively as they were found to be eligible for inclusion in
the study. Randomisation was performed in a 1:1 allocation, stratied by site and gender in blocks to
either HBOT + best practice or best practice. The randomisation sequence was computer-generated using
an internet-based application; RANDOMIZE.NET, which secured that the outcome of the randomisation,
i.e., treatment group, were masked until the time point at which each subject was to be randomised.
Procedures
All patients in the HBOT group received 60 minutes at 2.4 ATA with 10 minutes
compression/decompression time with one air-break making the total treatment time approximately 80
minutes.
The full procedure list is previously published and available online(23). Within 24 hours of randomisation,
subjects allocated to HBOT received their rst treatment. Subjects then received a maximum of 5
treatments within seven days of randomisation. NEWS and PFI was recorded three times a day for both
groups, the HBOT group also had NEWS/PFI recorded before/after the treatment. Both groups were
managed according to best practice including nasal high ow oxygen, non-invasive ventilation, and trial
of awake prone positioning at the discretion of the treating physician. Subjects received medical
treatment including corticosteroids and low molecular weight heparin (LMWH). All concomitant
medications including normobaric oxygen were recorded. Adverse events (AE) were recorded and
evaluated according to protocol as AE or serious adverse events (SAE) and graded as mild, moderate, or
severe. Causality in relation to hyperbaric oxygen (HBO2) was also assessed. Staff safety was evaluated
by any reports made through the hospitals reporting system for reporting negative events. The mean
oxygen dose was recorded 3 times daily for the past 8 hours and the cumulative “Oxygen burden” was
calculated at daily basis and longitudinally over the course of the trial (30 days). Oxygen is considered
toxic for healthy subjects at any FiO2 > 0.5 ≅ PO2 > 50 kPa(24). Oxygen toxicity is traditionally calculated

Page 6/19
as units of pulmonary toxic dose (UPTD) and with repeated exposures expressed as cumulative
pulmonary toxic dose (CPTD). We used a simplied calculation derived from the traditional equation:
UPTD = t × [0.5/(PO2 − 0.5)]−5/6, with PO2 in ATA and time in minutes(25). The simplied equation for
clinical use is dened as: UPTDICU= t × (FiO2 − 0.5) and summarised the daily UPTD values is used to
calculate the total “dose” of oxygen above FiO2 0.5 (CPTDICU). In the HBOT group one hour of mean
oxygen was replaced with the “HBOT dose”.
To reduce the bias, data were monitored by an independent monitor that checked source data for all
adverse events and selected source data for explanatory endpoints according to the monitoring plan.
Outcomes
The results of the primary and main secondary endpoints have not been evaluated. The trial protocol
including safety endpoints is previously described and available online(23).
Safety endpoints are the number of subjects, proportion of subjects and number of events of
AE/SAE/SADR, mean change in PaO2/FiO2 before and after HBOT compared with mean variance in
PaO2/FiO2 in the control group day 1 to day 7 mean change in NEWS before and after HBOT compared
with mean change in daily NEWS in the control group day 1 to day 7. Any negative events in staff
associated with treatment of subjects was also registered.
Exploratory outcomes associated with safety analysed in
the interim analysis:
(1) Mean oxygen dose per day including HBO2 and cumulative pulmonary oxygen toxicity expressed as
units of oxygen pulmonary toxicity dose (UPTD) and Cumulative pulmonary toxicity dose (CPTD) from
day 1 to day 30.
(2) Number of secondary infections, number of events and patients from day 1 to day 30.
(3) Diagnosed PE needing treatment, number of events and patients from day 1 to day 30.
Statistical analysis
Safety analyses were performed on the Safety population. (CONSORT ow diagram, Fig.1.)
Safety endpoints are adverse events (AE), vital parameters (NEWS) and oxygenation (PFI). Statistical
analysis for the NEWS and PFI scores was performed with the analysis of covariance (ANCOVA)
including baseline levels as a covariate, and treatment as a xed factor in the models. The null
hypothesis was no difference between the treatment groups. Tests were two-sided with a type-I error rate
of 0.05, where p-value < 0.05 is regarded as statistically signicant. There was no adjustment for
multiplicity as the safety endpoints and corresponding results are regarded as exploratory. Analysis was
performed on the safety population with observed data.

Page 7/19
Descriptive analysis of the number and percentage of patients reporting AEs, and the number of AEs
reported are presented. SAEs are also presented in separate tabulations. The events are tabulated by
system organ class and preferred term.
All continuous safety variables, such as age, body mass index (BMI), days with symptoms, number of risk
factors are described using summary statistics.
All categorical variables, such as ethnicity, smoking habits and, are summarised using frequencies and
percentages.
Results
Of the 54 patients assessed for eligibility, 31 subjects were randomised, 16 in the HBOT group and 15 in
the Best practice group. One patient was excluded from analysis due to withdrawal of consent before the
rst treatment. One patient was excluded from the control group due to a negative SARS-CoV-2 test and
was positive for adenovirus. Three subjects died in the trial; two in the HBOT group and one in the Control
group. Primary outcome (ICU admission) has not yet been analyzed, the cutoff date for inclusion in this
interim safety analysis was October 1, 2021. Gender distribution, age, weight, degree of disease was
evenly distributed between the groups. The subjects had moderate to severe ARDS at inclusion; and
groups were balanced between best practice vs HBOT at inclusion (Mean and SD). NEWS 5.4 (1.7) vs 5.3
(2.0), PaO2/PFiO217.3 (6.4) vs 14.0 (3.5) kPa (Table 1.)
Table 1. Baseline characteristics

Page 8/19
Baseline variable HBOT+ Best practice
N=15
Best practice
N=15
Age 64.4 (11.1) 63.3 (8.2)
Male gender 9 (60 %) 8 (53.3%)
Caucasian ethnicity 14 (93.3%) 15 (100%)
BMI 29.2 (4.4) 29.2 (5.0)
Number of risk factors 2.93 (0.96) 3.13 (1.06)
Smoker (every day) 1 (6.7%) 0 (0%)
Former smoker 5 (33.3%) 5 (33.3%)
Never smoker 9 (60%) 10 (66.7%)
Time since initial symptoms 9.93 (3.58) 11.67 (3.62)
NEWS at randomisation 5.3 (2.0) 5.4 (1.7)
PaO2/FiO2 at randomisation 14.0 (3.5) 17.3 (6.4)
Table 1.
Baseline characteristics expressed as Mean (SD) and Number and (Percentage) Full Analysis Set
(FAS) population
Those who received HBOT had a greater numerical improvement in NEWS and PFI days 7, 14 and 30. The
changes were not statistically signicant except for change in PFI day 14 (p=0.023) but this was not a
predened safety endpoint. (Figure 2.).
Adverse events (AE) were most commonly reported, in terms of hypoxia, slightly different distribution, and
grade of adverse events. A total of 95 AEs were registered; of the 23 Serious Adverse Events (SAE), nine
(in six subjects) were in the HBOT group and 14 (in six subjects) in the Control group. One SAE (hypoxia)
coincided with HBO2 treatment and lead to intubation within one hour after HBOT, hence possibly related
to HBO2 and was assessed as Serious Adverse Drug Reaction (SADR) even though the ICU admission
was planned before the treatment. (Table 2).

Page 9/19
Table 2. Show the number subjects, number of AEs and fraction of subjects with AEs, Serious AEs and
Severe AEs and type of SAE

There were no reported negative events in staff associated with the treatment of the subject (e.g., contact
with aerosol from the subject).
We analysed some predened exploratory endpoints associated with safety.
1.Oxygen toxicity and total oxygen dose in a subgroup (all patients from Karolinska University Hospital,
n=20). The mean UPTD was not signicantly higher in the HBOT group despite additional HBO2 (data not
shown). The cumulative oxygen burden expressed as CPTDICU was statistically not signicantly different
between the groups (Mean and SEM) HBOT 1618 (597) and Best Practice 1724 (716), (P=0.882) (Figure
3A). There was a trend towards faster recovery in terms of days with supplemental oxygen in the HBOT
group (P=0.318) (Figure 3B).
Figure 3Aand 3B.Cumulative oxygen burden and days with supplemental oxygen
2. Secondary infections: Two ventilatory associated pneumonias (VAPs), two bacteraemia/sepsis and
one abscess in m. obturatorious in the control group. One urinary tract infection in the HBOT group.
3. Thrombotic events: One patient in the HBOT group had a small pulmonary embolism.
4. Barotrauma: One subject in the control group had a pneumothorax. Five subjects had
pneumomediastinum, four in the control group and one in the HBOT group.
A complete list of adverse events with Medical Dictionary for Regulatory Activities (MedDRA) coding
including system organ class (SOC), the preferred term (PT) and code is available as supplementary
material, in Online resource 1.

Page 10/19
Discussion
This interim report focuses on the safety endpoints and the exploratory endpoints related to safety; more
specically on oxygen toxicity and barotrauma since these have been major concerns regarding HBOT for
COVID-19. There was a trend towards lower NEWS and higher PaO2/FiO2, with a statistically signicant
difference in PaO2/FiO2 at 14 days (Fig.2.). There was no statistically signicant difference in AE/SAE or
barotrauma between the groups, while numerically more AE/SAE and barotraumas in the control group.
Despite the added exposure of HBOT the rst 7 days, there was no difference in CPTDICU and a trend
towards lower probability of need for supplemental oxygen. While not powered for ecacy and the main
ecacy endpoints have not been analysed, based on the safety endpoints together with the exploratory
endpoints related to safety, our data show a trend towards benet for HBOT.
The SARS-CoV-2 pandemic has led to an unparalleled number of ARDS patients. To better target COVID-
19 patients at risk of mortality, a more pragmatic denition than the Berlin denition for COVID-19
induced ARDS has been suggested; including Nasal High Flow Oxygen (NHFO) FiO2 > 0.35 and
≥20L/min, accepting a 5–14 day window and unilateral opacities (26). There is an ongoing debate
whether COVID-19 induced acute respiratory failure (C-ARF) should be treated as a separate entity from
traditional ARDS(27, 28). A recent Delphi expert consensus statement agreed on that the pathophysiology
of C-ARF is similar to that of ARDS (29).
Some major concerns with HBOT in ARDS patients are risk of barotrauma, absorption atelectasis and
POT. These risks are also arguments against using HBOT in Severe COVID-19(12). Due to early fear of
oxygen toxicity the treatment protocol has a wide range at the discretion of the treating physician (1.6–
2.4 ATA for 30–60 minutes with 5–10 minutes compression time and 5–10 minutes decompression
time). Barotrauma during mechanical ventilation is well known to all intensivists and there is a strong
consensus regarding pressure and volume limits in ARDS(29).
The majority of subjects in our trial were on non-invasive mechanical ventilation (NIV) or nasal high ow
oxygen (NHFO) but we did not treat any patients on invasive mechanical ventilation with HBO2. There
was no evidence of increased barotrauma in the HBOT group; on the contrary pneumothorax and
pneumomediastinum were more frequent in the control group. It has previously been suggested that
HBO2 can be safely administered in mechanically ventilated patients with ARDS in a prospective
observational study, but it has not been evaluated in a randomised controlled trial(30).
Ventilation with FiO2 1.0 is suggested to cause absorption atelectasis(31). It is debated whether
absorption atelectasis is clinically relevant in critically ill patients with an already high fraction of inspired
oxygen (FiO2)(32). There was no difference in number of AEs related to hypoxia between the groups. Only
one of these occurred within 6 hours of HBOT, hence evaluated as possibly related to the HBO2 treatment
and nally assessed as a SADR. Our data does not suggest that there was an increased risk of
atelectasis post HBO2 treatment. We assessed that the fear of desaturation due to absorption atelectasis
was greatly overrated since there was only one SAE (hypoxia leading up to intubation) that was possibly

Page 11/19
related to the HBOT. This specic subject was on NIV with FiO2 0.8 before the HBOT and the ICU
admission was planned before the treatment. To evaluate oxygen toxicity, we recorded the mean daily
dose of oxygen, which is rarely done in clinical practice.
Toxic effects on the lungs caused by oxygen were discovered more than a century ago(33). POT can be
divided into two phases, an early exudative and a late proliferative phase(34). These two phases have
been experienced clinically by most intensivists treating ARDS patients, where the early phase is
reversible and the latter leading to brosis is irreversible if the oxygen fraction cannot be lowered below a
toxic dose(35). Even though POT is well-known, some of the damage caused in the ICU may be
iatrogenic. Supplemental oxygen has also been associated with negative effects on many other organs
which have implications for the critically ill patient(36). Most intensivist would avoid FiO2 1.0(29) but the
POT limit of FiO2 > 0.5 that is well known in diving- and hyperbaric medicine is rarely discussed in
intensive care medicine and oxygen toxicity is rather targeted at PaO2 or peripheral saturation (SpO2)(13).
Efforts have been made to establish a dose-response equation for the toxic effects but there is not a
linear correlation(37). The equation is extrapolated from mice, tested on healthy individuals and there are
numerous other factors apart from the concentration*time integral(37, 38). Despite attempts to measure
or better predict POT the gold standard for healthy divers is still calculated with the traditional equation
(see method)(25). For clinical use in the critical care setting, we suggest the simplied equation:
UPTDICU=t ×(FiO2 − 0.5) which can be easily applied by mental arithmetic to get an estimate of POT. It is
well known that intermittent reductions of oxygen, “air-breaks”, reduce harm and may even be
benecial(39, 40). Daily UPTDICU values can easily be collected and summarised as CPTDICU to evaluate
the risk of pulmonary toxicity over time.
Even though the UPTD is not exactly calculated it provides a rough estimate that can be used when
evaluating POT in clinical trials. We further calculated the cumulative oxygen burden since it is unknown
whether any FiO2 above normal air (21% O2) is toxic to injured lungs. If UPTDICU is calculated daily, we
speculate that it may also be benecial to use HBO2 in patients on invasive mechanical ventilation to
reduce inammation and to reduce cumulative oxygen burden.
Limitations
This was an open-label trial, neither the patients nor investigators were blinded to the allocated treatment.
Both groups were treated according to “best-practice” by independent staff with the addition of maximum
ve HBO treatments for subjects allocated to HBO + best practice. Due to logistical reasons, it was not
possible to conduct a single-blinded trial and since the safety of HBO for this indication was not
previously evaluated, we chose the open-label design. In this multicenter trial, subjects were enrolled in
different phases of the pandemic. The rst 8 subjects were enrolled before corticosteroids were
considered “best practice”; this may have affected the outcome of these subjects negatively. On the other
hand corticosteroids are still debated in ARDS due to multiple negative effects such as hyperglycaemia,
infections and weakness(41). We found a similar incidence of SAEs in these subjects compared to

Page 12/19
subjects included later and they were equally distributed between the two groups, but interestingly
secondary infections were more frequent in the control group, they all received concomitant
corticosteroids. Most subjects were included during the second and third wave and none of the subjects
were fully vaccinated, something that may have affected the outcome of this analysis and may affect
later ecacy analysis when subjects from subsequent waves are included. Due to the small sample size
of this interim analysis the results are prone to both type I and II errors.
Conclusion
Based on the interim safety analysis of our randomised controlled trial, we propose that HBOT is well
tolerated and can be safely used as an intervention for critically ill patients with moderate to severe ARDS
induced by COVID-19. We speculate that HBOT may be useful in ARDS from other cause even in
mechanically ventilated patients to reduce cumulative oxygen burden. Larger randomised controlled trials
are warranted to conrm safety and evaluate ecacy.
Declarations
Ethical Approval and Consent to participate
The trial was approved by National/Regional independent ethics committees in Sweden/Germany and
approved by the medical product agencies in Sweden/Germany. The trial is conducted according to
principles of the Declaration of Helsinki and complies with the International Conference on
Harmonization of technical requirements for registration of pharmaceuticals for human use, Good
Clinical Practice (ICH-GCP) . All patients have received oral and written information and signed an
Informed Consent Form before any trial specic activities takes place.
Trial registration:NCT04327505(March 31, 2020) and EudraCT 2020-001349-37 (April 24, 2020).
Funding
The trial was funded by The Swedish Research Council (Vetenskapsrådet) grant(KBF 2019–00446)
made available by redirecting funds to COVID-19 research originally awarded to Kenny Rodriguez-
Wallberg.Internal funding from PeterLindholm, dept of physiology and pharmacology, KI.PL is
supported by the Ted and Michelle Gurnee Endowed Chair for Hyperbaric medicine research at University
of California San Diego.Grants from Konung Gustav V:s och Drottning Victorias frimurarestiftelse and
von Kantzow stiftelse was made availible by Sergiu-Bogdan Catrina.
Author contributions
AK conceptualized the hypothesis, trial design and wrote the protocol together with PL.KR-W, JD and S-
BC contributed with information to the protocol and IRB/MPA applications.AK is coordinating
investigator for the trial and principal investigator at Karolinska University Hospital. PL is sponsor

Page 13/19
representative. JD is principal investigator at Blekingesjukhuset, Karlskrona. AH, SA-E and EB are sub-
investigators and have been involved in patient management and revisions of protocols and standard
operations procedures. Trial material,including electronic case report form, patient information, standard
operations procedures etc. was prepared by AK and JD and revised by PL and AH. Data
collectionwasperformed byAK, JD, AH, SA-E, EB, LAH, EH, ADJ-N, and LL. Dataanalysis was performed
by AK, SA-E and JK.The manuscript wasdraftedbyAKand all authors commented on previous versions
and approved the nal manuscript. PL and KR-Wequally contributedas senior authors.AK is
corresponding author for this work and attests that all listed authors meet authorship criteria and that no
others meeting the criteria have been omitted.
Consent for publication
All authorshaveapprovedthe manuscript and give their consent for submission and publication.
Disclosures and declarations
All authorsdeclare that they have no known competing nancial interests or personal relationships that
could have appeared to inuence the work reported in this paper.
Acknowledgements
We acknowledgeall hospital staff at Karolinska University Hospital (KUH) and Blekingesjukhuset
Karlskrona (BSK) involved in the management of the patients. We thank all subjects that volunteered for
the trial. The director of Intensive care, KUH, Björn Persson, and head of department Physiology and
Pharmacology at Karolinska Institutet (KI), Håkan Westerblad for supporting this project. The research
nurses at kliniska forskningsenheten (KFE) at KUH, staff at Studiecenter Karolinska for setting up the
laboratory manual and handling blood samples.The members of the independent Data Safety
Monitoring Board: Magnus Nord (Chair), Miklos Lipcsey and Anders Öwall for reviewing data. PhD
student Allan Zhao and medical student Pontus Hedberg for help with data collection.
Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author
on reasonable request.
References
1. Ferguson ND, Fan E, Camporota L, Antonelli M, Anzueto A, Beale R, Brochard L, Brower R, Esteban A,
Gattinoni L, Rhodes A, Slutsky AS, Vincent JL, Rubenfeld GD, Thompson BT, Ranieri VM. The Berlin
denition of ARDS: an expanded rationale, justication, and supplementary material. Intensive care
medicine 2012; 38: 1573–1582.
2. Novak S, Drenjancevic I, Vukovic R, Kellermayer Z, Cosic A, Tolusic Levak M, Balogh P, Culo F, Mihalj
M. Anti-Inammatory Effects of Hyperbaric Oxygenation during DSS-Induced Colitis in BALB/c Mice

Page 14/19
Include Changes in Gene Expression of HIF-1alpha, Proinammatory Cytokines, and Antioxidative
Enzymes.
Mediators Inamm
2016; 2016: 7141430.
3. Salhanick SD, Belikoff B, Orlow D, Holt D, Reenstra W, Buras JA. Hyperbaric oxygen reduces
acetaminophen toxicity and increases HIF-1alpha expression. Acad Emerg Med 2006; 13: 707–714.
4. Choudhury R. Hypoxia and hyperbaric oxygen therapy: a review. Int J Gen Med 2018; 11: 431–442.
5. Bennett MH, Feldmeier J, Hampson NB, Smee R, Milross C. Hyperbaric oxygen therapy for late
radiation tissue injury. The Cochrane database of systematic reviews 2016; 4: CD005005.
. Oscarsson N, Muller B, Rosen A, Lodding P, Molne J, Giglio D, Hjelle KM, Vaagbo G, Hyldegaard O,
Vangedal M, Salling L, Kjellberg A, Lind F, Ettala O, Arola O, Seeman-Lodding H. Radiation-induced
cystitis treated with hyperbaric oxygen therapy (RICH-ART): a randomised, controlled, phase 2–3 trial.
The lancet oncology 2019; 20: 1602–1614.
7. Kirby JP. Hyperbaric Oxygen Therapy and Radiation-Induced Injuries. Mo Med 2019; 116: 198–200.
. Dulai PS, Raffals LE, Hudesman D, Chiorean M, Cross R, Ahmed T, Winter M, Chang S, Fudman D,
Sadler C, Chiu EL, Ross FL, Toups G, Murad MH, Sethuraman K, Holm JR, Guilliod R, Levine B, Buckey
JC, Jr., Siegel CA. A phase 2B randomised trial of hyperbaric oxygen therapy for ulcerative colitis
patients hospitalised for moderate to severe ares. Aliment Pharmacol Ther 2020.
9. Kirby JP. Hyperbaric Oxygen Indications: Diabetic Foot Ulcers and Intractable Management. Mo Med
2019; 116: 188–191.
10. Devaney B, Frawley G, Frawley L, Pilcher DV. Necrotising soft tissue infections: the effect of
hyperbaric oxygen on mortality. Anaesth Intensive Care 2015; 43: 685–692.
11. Bartek J, Jr., Jakola AS, Skyrman S, Forander P, Alpkvist P, Schechtmann G, Glimaker M, Larsson A,
Lind F, Mathiesen T. Hyperbaric oxygen therapy in spontaneous brain abscess patients: a population-
based comparative cohort study. Acta Neurochir (Wien) 2016; 158: 1259–1267.
12. Moon RD, Weaver LK. Hyperbaric oxygen as a treatment for COVID-19 infection? Undersea Hyperb
Med 2020; 47: 177–179.
13. Singer M, Young PJ, Laffey JG, Asfar P, Taccone FS, Skrifvars MB, Meyhoff CS, Radermacher P.
Dangers of hyperoxia. Critical care (London, England) 2021; 25: 440.
14. Sadler C, Alvarez Villela M, Van Hoesen K, Grover I, Lang M, Neuman T, Lindholm P. Diving after
SARS-CoV-2 (COVID-19) infection: Fitness to dive assessment and medical guidance. Diving and
hyperbaric medicine: the journal of the South Pacic Underwater Medicine Society 2020; 50: 278–
287.
15. Paganini M, Bosco G, Perozzo FAG, Kohlscheen E, Sonda R, Bassetto F, Garetto G, Camporesi EM,
Thom SR. The Role of Hyperbaric Oxygen Treatment for COVID-19: A Review. Adv Exp Med Biol 2021;
1289: 27–35.
1. Zhong XT, X; Tang, Y; Chen, R;. The effect of hyperbaric oxygen therapy on hypoxia in patients with
severe new coronavirus pneumonia: the rst report. Chinese Journal of Nautical Medicine and
Hyperbaric Medicine 2020.

Page 15/19
17. Thibodeaux K, Speyrer M, Raza A, Yaakov R, Serena TE. Hyperbaric oxygen therapy in preventing
mechanical ventilation in COVID-19 patients: a retrospective case series. J Wound Care 2020; 29: S4-
S8.
1. Gorenstein SA, Castellano ML, Slone ES, Gillette B, Liu H, Alsamarraie C, Jacobson AM, Wall SP,
Adhikari S, Swartz JL, McMullen JJS, Osorio M, Koziatek CA, Lee DC. Hyperbaric oxygen therapy for
COVID-19 patients with respiratory distress: treated cases versus propensity-matched controls.
Undersea Hyperb Med 2020; 47: 405–413.
19. Cannellotto M, Duarte M, Keller G, Larrea R, Cunto E, Chediack V, Mansur M, Brito DM, Garcia E, Di
Salvo HE, Verdini F, Dominguez C, Jorda-Vargas L, Roberti J, Di Girolamo G, Estrada E. Hyperbaric
oxygen as an adjuvant treatment for patients with COVID-19 severe hypoxaemia: a randomised
controlled trial.
Emergency medicine journal: EMJ
2021.
20. Oliaei S, SeyedAlinaghi S, Mehrtak M, Karimi A, Noori T, Mirzapour P, Shojaei A, MohsseniPour M,
Mirghaderi SP, Alilou S, Shobeiri P, Azadi Cheshmekabodi H, Mehraeen E, Dadras O. The effects of
hyperbaric oxygen therapy (HBOT) on coronavirus disease-2019 (COVID-19): a systematic review. Eur
J Med Res 2021; 26: 96.
21. Kjellberg A, De Maio A, Lindholm P. Can hyperbaric oxygen safely serve as an anti-inammatory
treatment for COVID-19? Medical Hypotheses 2020; 144.
22. De Maio A, Hightower LE. COVID-19, acute respiratory distress syndrome (ARDS), and hyperbaric
oxygen therapy (HBOT): what is the link? Cell Stress Chaperones 2020: 1–4.
23. Kjellberg A, Douglas J, Pawlik MT, Kraus M, Oscarsson N, Zheng X, Bergman P, Franberg O, Kowalski
JH, Nyren SP, Silvanius M, Skold M, Catrina SB, Rodriguez-Wallberg KA, Lindholm P. Randomised,
controlled, open label, multicentre clinical trial to explore safety and ecacy of hyperbaric oxygen for
preventing ICU admission, morbidity and mortality in adult patients with COVID-19. BMJ Open 2021;
11: e046738.
24. Winter PM, Smith G. The toxicity of oxygen. Anesthesiology 1972; 37: 210–241.
25. Wingelaar TT, Brinkman P, de Vries R, van Ooij PAM, Hoencamp R, Maitland-van der Zee AH,
Hollmann MW, van Hulst RA. Detecting Pulmonary Oxygen Toxicity Using eNose Technology and
Associations between Electronic Nose and Gas Chromatography-Mass Spectrometry Data.
Metabolites 2019; 9.
2. Brown SM, Peltan ID, Barkauskas C, Rogers AJ, Kan V, Gelijns A, Thompson BT. What Does Acute
Respiratory Distress Syndrome Mean during the COVID-19 Pandemic? Ann Am Thorac Soc 2021; 18:
1948–1950.
27. Bos LDJ, Sjoding M, Sinha P, Bhavani SV, Lyons PG, Bewley AF, Botta M, Tsonas AM, Serpa Neto A,
Schultz MJ, Dickson RP, Paulus F, group PR-Cc. Longitudinal respiratory subphenotypes in patients
with COVID-19-related acute respiratory distress syndrome: results from three observational cohorts.
The lancet Respiratory medicine 2021; 9: 1377–1386.
2. Lascarrou JB. COVID-19-related ARDS: one disease, two trajectories, and several unanswered
questions. The lancet Respiratory medicine 2021; 9: 1345–1347.

Page 16/19
29. Nasa P, Azoulay E, Khanna AK, Jain R, Gupta S, Javeri Y, Juneja D, Rangappa P, Sundararajan K,
Alhazzani W, Antonelli M, Arabi YM, Bakker J, Brochard LJ, Deane AM, Du B, Einav S, Esteban A, Gajic
O, Galvagno SM, Jr., Guerin C, Jaber S, Khilnani GC, Koh Y, Lascarrou JB, Machado FR, Malbrain M,
Mancebo J, McCurdy MT, McGrath BA, Mehta S, Mekontso-Dessap A, Mer M, Nurok M, Park PK,
Pelosi P, Peter JV, Phua J, Pilcher DV, Piquilloud L, Schellongowski P, Schultz MJ, Shankar-Hari M,
Singh S, Sorbello M, Tiruvoipati R, Udy AA, Welte T, Myatra SN. Expert consensus statements for the
management of COVID-19-related acute respiratory failure using a Delphi method.
Critical care
(London, England)
2021; 25: 106.
30. Bessereau J, Aboab J, Hullin T, Huon-Bessereau A, Bourgeois JL, Brun PM, Chevret S, Annane D.
Safety of hyperbaric oxygen therapy in mechanically ventilated patients. Int Marit Health 2017; 68:
46–51.
31. O'Brien J. Absorption atelectasis: incidence and clinical implications. AANA J 2013; 81: 205–208.
32. Joyce CJ, Baker AB. What is the role of absorption atelectasis in the genesis of perioperative
pulmonary collapse? Anaesth Intensive Care 1995; 23: 691–696.
33. Smith JL. The pathological effects due to increase of oxygen tension in the air breathed. The Journal
of physiology 1899; 24: 19–35.
34. Wingelaar TT, van Ooij PAM, van Hulst RA. Oxygen Toxicity and Special Operations Forces Diving:
Hidden and Dangerous. Front Psychol 2017; 8: 1263.
35. Michalski JE, Kurche JS, Schwartz DA. From ARDS to pulmonary brosis: the next phase of the
COVID-19 pandemic? Transl Res 2021.
3. Kallet RH, Branson RD. Should Oxygen Therapy Be Tightly Regulated to Minimize Hyperoxia in
Critically Ill Patients? Respiratory care 2016; 61: 801–817.
37. Arieli R, Yalov A, Goldenshluger A. Modeling pulmonary and CNS O(2) toxicity and estimation of
parameters for humans. J Appl Physiol (1985) 2002; 92: 248–256.
3. Wingelaar TT, Brinkman P, Hoencamp R, van Ooij PA, Maitland-van der Zee AH, Hollmann MW, van
Hulst RA. Assessment of pulmonary oxygen toxicity in special operations forces divers under
operational circumstances using exhaled breath analysis. Diving and hyperbaric medicine: the
journal of the South Pacic Underwater Medicine Society 2020; 50: 2–7.
39. Chavko M, Mahon RT, McCarron RM. Mechanisms of protection against pulmonary hyperbaric O(2)
toxicity by intermittent air breaks. European journal of applied physiology 2008; 102: 525–532.
40. Zhang L, Zhang Y, Wang Z, Chen Y, Li R. Intermittent hyperbaric oxygen exposure mobilizing
peroxiredoxin 6 to prevent oxygen toxicity. J Physiol Sci 2019; 69: 779–790.
41. Hensley MK, Sjoding MW, Prescott HC. COUNTERPOINT: Should Corticosteroids Be Routine
Treatment in Early ARDS? No. Chest 2021; 159: 29–33.
Figures

Page 17/19
Figure 1
CONSORT ow diagram

Page 18/19
Figure 2
Changes from baseline in NEWS and PFI day 7, day 14 and day 30

Page 19/19
Figure 3
3A and B. Cumulative oxygen burden expressed as CPTDICU day 1 to day 30 and Kaplan-Meier curve
describing the cumulative probability of need for supplemental oxygen day 1 to day 30
.
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.
Onlineresource1.xlsx
TakeHomemessage.png