Hyperbaric Oxygen Therapy for Pneumonia: What the Evidence Actually Shows

Hyperbaric Oxygen Therapy for Pneumonia: What the Evidence Actually Shows

Breathing pure oxygen inside a pressurised chamber has been used for decades in conditions ranging from decompression sickness to chronic wound healing. More recently — driven largely by the COVID-19 pandemic — researchers have turned their attention to whether hyperbaric oxygen therapy (HBOT) could help patients with severe pneumonia. The question is reasonable: when lungs are flooded with inflammation and can't deliver enough oxygen, could bypassing the usual pathway help? Here's what the clinical evidence actually says.

Article Summary: HBOT has a small but biologically plausible evidence base as an adjunctive therapy for COVID-19 pneumonia. Two small randomised trials and a living systematic review suggest it may improve oxygenation and shorten hospital stays, but a meaningful mortality benefit has not been established. No clinical evidence exists for HBOT in bacterial pneumonia or any non-COVID viral pneumonia (influenza, RSV, MERS). The intervention remains investigational.

The Problem: When Lungs Can't Keep Up

Pneumonia is the leading infectious cause of death globally, responsible for approximately 4 million deaths each year. When pneumonia progresses to acute respiratory distress syndrome (ARDS) — as it did in 10–23% of COVID-19 hospitalisations — the lungs become so inflamed and fluid-filled that standard oxygen delivery through a mask or nasal cannula may not be enough.

During the worst of the pandemic, mortality rates for mechanically ventilated COVID-19 ARDS patients exceeded 40% in many centres. For patients deteriorating on high-flow oxygen but not yet intubated, clinicians faced a narrow window with limited options. This is the gap that HBOT was proposed to fill.

The core challenge in severe pneumonia is not just infection — it's the inflammatory cascade that follows. The immune system's response can become self-destructive, damaging alveolar tissue, flooding air spaces with fluid, and creating a cycle of worsening hypoxia that standard oxygen therapy struggles to break.

What Is HBOT and Why Might It Help?

Hyperbaric oxygen therapy delivers 100% oxygen at pressures above normal atmospheric pressure — typically 1.5 to 2.4 atmospheres absolute (ATA). Under these conditions, dissolved oxygen in blood plasma increases dramatically — roughly 1.5 mL O₂ per decilitre for each additional ATA — independent of haemoglobin. In practical terms, this means oxygen can reach tissues even when the normal lung-to-blood gas exchange is severely compromised.

For pneumonia specifically, the proposed benefits operate through three mechanisms:

How HBOT May Work in Pneumonia: Four Mechanisms

1. Direct Oxygenation Bypass At 2.4 ATA, plasma-dissolved oxygen reaches approximately 6 mL/dL — enough to sustain basal metabolic demands without haemoglobin. In patients with consolidated or fluid-filled lungs, this provides a route around the broken gas-exchange barrier. Multiple COVID-19 case reports documented immediate improvements in SpO₂ and PaO₂/FiO₂ ratio during and after HBOT sessions.

2. Cytokine Storm Suppression Severe pneumonia — particularly COVID-19 — involves a dysregulated "cytokine storm" with elevated TNF-α, IL-6, IL-1β, CRP, and ferritin. HBOT at therapeutic pressures suppresses NF-κB-mediated transcription of these pro-inflammatory cytokines, reduces neutrophil adhesion molecule expression, and shifts macrophage polarisation from pro-inflammatory (M1) toward anti-inflammatory (M2) phenotypes. One small RCT (Siewiera 2022) documented significant reductions in CRP, ferritin, and LDH alongside increased CD3 T-cell counts in HBOT-treated patients.

3. Immune Cell Restoration In hypoxic infected tissue, neutrophils and macrophages lose their oxidative burst capacity — the oxygen-dependent killing mechanism that is their primary weapon against pathogens. By restoring tissue oxygen tension, HBOT reactivates this antimicrobial function. This is particularly relevant in aspiration pneumonia where anaerobic bacteria thrive in low-oxygen environments.

4. Pressure-Dependent Anti-Inflammatory Action Animal studies (Chu 2007) have shown that hyperbaric pressure — not merely high oxygen concentration — is required for TNF-α suppression. Normobaric hyperoxia (100% O₂ at 1 ATA) did not produce the same anti-inflammatory effect. This pressure-dependent mechanism, supported by additional animal data showing HO-1/Nrf2 pathway activation and aquaporin restoration, suggests HBOT acts through specific molecular pathways rather than simple oxygen delivery.

What the Clinical Research Shows

The evidence base for HBOT in pneumonia is recent, small, and restricted almost entirely to COVID-19. Here is what exists.

The Landmark Trial

The most methodologically rigorous trial was published by Kjellberg et al. in Pulmonary Pharmacology & Therapeutics (2024). Twenty-three COVID-19 moderate ARDS patients were randomised to HBOT (2.4 ATA, 80 minutes, 5 sessions within 7 days) plus standard care or standard care alone. Seventeen were analysed. The HBOT group had a significantly shorter hospital stay — 16 days versus 26 days (p=0.045) — with improved oxygenation scores and NEWS clinical assessments throughout admission. RNA-sequencing revealed 791 differentially expressed genes in HBOT patients (versus 46 in controls) at Day 7, with a novel endoplasmic reticulum stress transcriptomic signature.

Key Clinical Evidence at a Glance

Study Finding
Kjellberg 2024 (RCT, n=17 analysed) Hospital stay 16 vs 26 days (p=0.045); improved PaO₂/FiO₂ and NEWS scores; 791 DEGs at Day 7
Siewiera 2022 (RCT, n=28) 0 deaths in HBOT vs 3/14 (21%) in controls; decreased CRP, ferritin, LDH; increased CD3 T-cells
Boet 2022 (Living SR, 8 studies, n=224) All 8 studies reported improved outcomes; no serious adverse events; meta-analysis not feasible due to heterogeneity
Jansen 2022 (Observational, n=36) Unadjusted mortality 13% vs 50%; after adjustment for confounders, HR 0.48 (p=0.42) — no significant difference
Palaniappan 2022 (Single-arm, n=50) 100% achieved PaO₂ ≥90 mmHg after 3 sessions — but no control group
Kjellberg 2023 (Safety RCT, n=31) Adverse events common (hypoxia most frequent); no excess serious AEs in HBOT group

A Critical Caution: The Jansen 2022 Lesson

One of the most important papers in this field is Jansen et al. (2022), which serves as a cautionary tale about confounding in observational data. Their unadjusted comparison showed HBOT patients had dramatically lower mortality — 13% versus 50%. But after adjusting for age and baseline hypoxia severity, the difference vanished (HR 0.48, p=0.42). The explanation: sicker, older patients were less likely to receive HBOT in the first place. This study is a powerful reminder that unadjusted mortality signals in HBOT pneumonia research should be interpreted with extreme caution.

What About Non-COVID Pneumonia?

This is where the evidence gap is most stark.

Bacterial Pneumonia (CAP and HAP)

No clinical studies of HBOT for bacterial community-acquired or hospital-acquired pneumonia exist — not RCTs, not cohort studies, not even case reports. One animal study (Hekimoglu Sahin 2011) showed HBOT reduced inflammatory markers in a rat aspiration pneumonia model, but this provides mechanistic rationale only.

Non-COVID Viral Pneumonia (Influenza, RSV, MERS, SARS-CoV-1)

Four dedicated PubMed searches conducted during this research run — targeting influenza, MERS-CoV, RSV, adenovirus, and broad viral respiratory infections — returned zero clinical papers of any design. Not a single case report of HBOT in influenza pneumonia, RSV pneumonia, or MERS has been published.

The absence is notable because the mechanistic rationale is genuinely shared across these pathogens: TNF-α elevation, HO-1 dysregulation, Nrf2 suppression, and aquaporin dysfunction are all documented in influenza, RSV, and coronavirus infections. Six pre-COVID animal studies (2002–2016) using LPS-induced acute lung injury models demonstrate that HBOT activates precisely the pathways relevant to viral pneumonia — but all used bacterial endotoxin, not actual viruses.

The bottom line: a mechanistic framework for potential benefit exists, but no clinician should extrapolate from COVID-19 data to other viral pneumonias. The pathogens cause different immunopathological profiles, the safety population is different, and the COVID-19 evidence itself has Low certainty.

Safety Profile

The safety data from HBOT in COVID-19 pneumonia is the most consistent finding in the field. Across approximately 200 patients in systematic reviews and trials, no serious HBOT-specific adverse events were reported:

  • No seizures from oxygen toxicity
  • No pneumothorax from barotrauma
  • No haemodynamic deterioration requiring cessation

Adverse events were common in the Kjellberg 2023 safety RCT (hypoxia was most frequent), but serious AEs were numerically fewer in the HBOT group.

Specific concerns for pneumonia patients include:

  • Pulmonary oxygen toxicity — a risk with prolonged high-pressure O₂, though the short protocols used (≤5 sessions) are well below established toxicity thresholds
  • Barotrauma — patients with bullae or severely damaged lung parenchyma face elevated pneumothorax risk during pressurisation
  • Infection control — COVID-era studies required full PPE inside chambers with enhanced decontamination, creating logistical constraints
  • Patient tolerance — critically ill patients may not tolerate chamber confinement, likely introducing selection bias in trials

The Evidence-Based Protocol

Parameter What Studies Used
Pressure 2.0–2.4 ATA (most rigorous trial: 2.4 ATA)
Duration 60–80 minutes per session
Sessions Up to 5 sessions within 7 days
Timing During moderate-to-severe ARDS, before intubation; patients deteriorating despite high-flow oxygen
Patient selection COVID-19 moderate ARDS with PaO₂/FiO₂ <200 mmHg plus ≥2 ICU risk factors (Kjellberg 2023 criteria)
Safety Confirmed tolerable for ≤5 sessions; no long-term safety data available

What the Research Doesn't Yet Tell Us

The most important gap is statistical power. No trial has been large enough to determine whether HBOT actually reduces mortality in pneumonia of any type. The most promising trial enrolled just 23 patients; the adjusted observational analysis found no significant survival benefit.

Protocol optimisation is entirely uncharted. Whether 2.0 ATA is as effective as 2.4 ATA, whether 3 sessions suffice or 5 are needed, and whether longer courses might improve outcomes — none of this has been studied in any head-to-head comparison. Session lengths ranged from 30 to 90 minutes across studies, with no dose-response data.

The restriction to COVID-19 is the field's largest limitation. All human evidence comes from a single disease during a specific pandemic context. Whether HBOT would help in influenza ARDS, RSV pneumonia, or severe bacterial CAP is completely unknown. The mechanistic case is plausible but has not been tested — even in animal models using actual respiratory viruses.

Finally, no study has reported outcomes beyond hospital discharge: no 30-day mortality, no 90-day follow-up, no quality of life data, no functional recovery metrics. Whether HBOT's in-hospital oxygenation benefits translate to meaningful long-term patient outcomes is unknown.

Bottom Line

HBOT is a biologically plausible adjunctive therapy for severe pneumonia, with a small but consistent signal from COVID-19 trials suggesting it may improve oxygenation and shorten hospital stays. The mechanism — plasma oxygen delivery independent of damaged lungs, plus anti-inflammatory and immunomodulatory effects — is well-characterised in both human and animal data.

However:

  • Mortality benefit is not established. The one study suggesting lower deaths was underpowered; the adjusted observational analysis found no significant effect.
  • All clinical evidence is from COVID-19 only. Bacterial pneumonia and non-COVID viral pneumonia (influenza, RSV, MERS) have zero clinical data.
  • The evidence is Low to Very Low certainty. Small, single-centre, unblinded trials cannot provide definitive answers.
  • The intervention is investigational. HBOT for pneumonia should not be considered standard care; it may be appropriate within a clinical trial setting.

For respiratory medicine professionals, the evidence warrants continued investigation — particularly multicentre RCTs powered for mortality in COVID-19 ARDS, and first-in-human trials for influenza and RSV ARDS, where the mechanistic rationale is strong but clinical validation is entirely absent.

Explore the Full Research

The full evidence synthesis for this topic is available in this research run:

Key References

  1. Kjellberg A et al. Hyperbaric oxygen therapy as an immunomodulatory intervention in COVID-19-induced ARDS: exploring clinical outcomes and transcriptomic signatures in a randomised controlled trial. Pulm Pharmacol Ther. 2024;87:102330.
  2. Siewiera J et al. Effectiveness of hyperbaric oxygen therapy in SARS-CoV-2 pneumonia: the primary results of a randomised clinical trial. J Clin Med. 2022;12(1):8.
  3. Boet S et al. Efficacy and safety of hyperbaric oxygen treatment to treat COVID-19 pneumonia: a living systematic review update. Diving Hyperb Med. 2022;52(2):126–135.
  4. Jansen D et al. Hyperbaric oxygen for COVID-19 patients with severe hypoxia prior to vaccine availability. Undersea Hyperb Med. 2022;49(3):295–305.
  5. Kjellberg A et al. COVID-19-induced acute respiratory distress syndrome treated with hyperbaric oxygen: interim safety report from a randomized clinical trial. J Clin Med. 2023;12(14):4850.
  6. Chu SJ et al. Effects of hyperbaric oxygen on endotoxin-induced acute lung injury in rats. Pulm Pharmacol Ther. 2007;20(2):162–166.
  7. Huang TY et al. Hyperbaric oxygen attenuation of lipopolysaccharide-induced acute lung injury involves heme oxygenase-1. Acta Anaesthesiol Scand. 2005;49(7):1022–1028.

Full reference list: 06-references.md

This article is for informational purposes only and does not constitute medical advice. The evidence summarised here reflects research available as of March 2026. HBOT for pneumonia is an investigational intervention; it should not be used outside a clinical trial setting without appropriate medical supervision. Individual responses may vary.

Back to blog