Hyperbaric oxygen therapy in 2026: what the newest evidence actually supports
A practitioner-facing review of the 2025–2026 hyperbaric oxygen therapy literature — effect sizes, session protocols, the null results that matter, and where the signal is strongest.
- A 2025 double-blind sham-controlled RCT in persistent post-brain-injury symptoms found a mean between-group improvement of 7.0 points on the Neurobehavioral Symptom Inventory (95% CI 1.7–12.3, p = 0.01) after 40 sessions.[1]
- A 2025 cohort of adults with post-concussion syndrome from childhood TBI showed significant gains in five of six cognitive domains, with medium-to-large effect sizes (r = 0.62–0.78) — a mean 23.6 years after injury.[2]
- A prospective long-COVID registry (n = 232) reported 61% of patients exceeding the minimal clinically important difference in quality of life at three months.[3]
- Counterweight: a 155-participant sham-controlled trial in older adults with type 2 diabetes and MCI found no cognitive benefit over sham — with sham numerically favoured at three months.[5]
- Across studies the working protocol converges on 40+ sessions, 90 minutes, 100% O₂ at 2.0 ATA with intermittent air breaks.
Hyperbaric oxygen therapy has spent two decades in an awkward position: mechanistically plausible, clinically established for a narrow set of indications, and aggressively marketed for everything else. The 2025–2026 literature is the first stretch in a while where the sham-controlled data, the registry data and the null results have all landed close enough together to draw a usable line between them.
This review is aimed at practitioners and at patients who read like practitioners. It covers what the newest trials found, at what dose, and with what confidence.
Considering HBOT as part of a clinical protocol? Peak Human designs pressure, duration and cadence around the indication — not a package.
Book an HBOT ConsultationThe mechanism, restated precisely
At 2.0 atmospheres absolute breathing 100% oxygen, plasma oxygen content rises roughly ten-fold over sea-level air. The therapeutic effect, though, is increasingly attributed not to the absolute oxygen load but to the hyperoxic–hypoxic paradox: the repeated cycling between hyperoxia at pressure and relative normoxia between sessions triggers the same HIF-1α-mediated transcriptional programs as hypoxia itself — angiogenesis, stem cell mobilisation, mitochondrial biogenesis — without the ischemic insult.
This matters clinically because it predicts dose dependence on session count, not session intensity. It explains why nearly every protocol showing an effect in the current literature runs 40 sessions or more, and why short courses tend to read as null.
The signal in the 2025–2026 data is not that oxygen heals. It's that repeated oxygen cycling, at sufficient session count, changes measurable outcomes in a subset of conditions.
Brain injury: the strongest new signal
Weaver et al., Scientific Reports (2025)
Adults with persistent symptoms after non-stroke brain injury (40 of 47 with TBI) received 40 HBO₂ or 40 sham sessions over 12 weeks. At the 13-week primary endpoint, NSI change was 10.6 ± 10.6 in the HBO₂ arm versus 3.6 ± 5.9 in sham — a mean difference of 7.0 (95% CI 1.7–12.3, p = 0.01). The HBO₂ group additionally improved on olfaction, anxiety, sleep difficulty and vestibular measures.[1]
Two caveats a clinician should hold onto. First, the sample is small — 47 analysed — so the confidence interval is wide and the point estimate should be treated as provisional. Second, both arms improved on depression, headache, PTSD symptoms and physical quality of life, which is a reminder of how much of the perceived benefit in open-label hyperbaric practice is attributable to structure, attention and time.
The second brain-injury dataset is retrospective but unusually clean in its inclusion criteria.
Shabi Shlifer et al., Frontiers in Neurology (2025)
Adults (mean age 31.7) who sustained TBI before age 17 and began HBOT after age 20 — a mean of 23.6 years post-injury. All received at least 40 sessions of 90 minutes at 100% O₂ / 2.0 ATA with air breaks. After Bonferroni correction, all cognitive domains except motor skills improved significantly, with effect sizes r = 0.62–0.78. Gains were independent of both time since injury and initial TBI severity.[2]
Standardised cognitive scores before vs. after HBOT
Fig. 1 — Adults with persistent post-concussion syndrome from childhood TBI (n = 26), ≥40 sessions at 2.0 ATA. All domains improved significantly except motor skills (adjusted p = 0.18). Data: Shabi Shlifer et al., Front Neurol 2025.[2]
Cognition beyond injury: a split result
A January 2026 study extended the cognitive question to patients who were in the chamber for entirely different reasons.
Betyna-Białek, Grobelska & Borkowska, Scientific Reports (2026)
Fifty-three patients aged 26–83 treated for idiopathic sudden sensorineural hearing loss (n = 23) or hard-to-heal chronic wounds (n = 30) underwent MMSE, verbal fluency, Trail Making, Rey Auditory Verbal Learning and Stroop testing before and after HBOT. Most cognitive parameters improved significantly in both groups. Notably, older age and greater baseline cognitive impairment predicted smaller gains.[4]
This is an uncontrolled pre/post design — practice effects on repeated neuropsychological testing are a real confound and the authors do not fully rule them out. Treat it as hypothesis-generating.
The necessary counterweight is a much larger and much more rigorous trial, and it did not go the way the field expected.
Benari et al., Alzheimer's & Dementia (2025)
155 older adults with type 2 diabetes and mild cognitive impairment were randomised to HBOT (n = 77) or sham (n = 78) between 2017 and 2023, with cognitive testing plus ASL-MRI and FDG-PET at baseline, 3, 6 and 12 months. All participants improved from baseline. But at three months, global cognition and executive function significantly favoured sham (p < 0.05), with no sustained between-group differences thereafter.[5]
This is the single most important recent result for anyone advising patients honestly. In a well-powered, well-blinded design, HBOT did not outperform sham for cognitive enhancement in a metabolically compromised older population. The plausible reading is that HBOT's cognitive effect is restorative in the presence of a specific injury or hypoperfused tissue, and not enhancing in its absence — but that reading is inference, not demonstrated fact.
Sham-controlled trials: HBOT vs. sham
Fig. 2 — Top: mean 13-week NSI improvement, brain injury RCT (higher = better).[1] Bottom: directional 3-month cognitive outcome in the T2D/MCI trial, where sham was numerically favoured (directional only).[5] Two rigorous trials, two different answers — indication appears to be the discriminating variable.
Our clinicians will tell you when the evidence supports HBOT for your indication — and when it doesn't.
Book an HBOT ConsultationLong COVID: real-world registry data
Lansdorp, van Berkel, Lalieu et al., Scientific Reports (2025)
Dutch multicentre registry of patients receiving off-label HBOT for long COVID. EQ-5D VAS rose from a mean of 37 at baseline to 46 immediately post-treatment and 50 at three months (both p < 0.001). Of 163 patients analysed, 99 (61%) exceeded the 7.5-point MCID; 16 (10%) deteriorated beyond it. Symptom-level improvement was largest for word-finding difficulty (73%), irritability (72%), memory problems (70%) and brain fog (68%).[3]
Quality of life (EQ-5D VAS) across the HBOT course
Fig. 3 — Long COVID registry, median 40 sessions (n = 138 analysed for the VAS trajectory). Gains were maintained and modestly extended at three-month follow-up. Data: Lansdorp et al., Sci Rep 2025.[3]
A registry is not a trial. There is no sham arm, participants self-selected into an off-label treatment they were often paying for, and 13% were classified as negative responders. A February 2026 systematic review covering 21 studies from January 2021 to October 2025 — including ten RCTs — reached a similarly hedged conclusion: HBOT appears to improve quality of life, fatigue, cognition and neuropsychiatric symptoms in long COVID, but indications, protocols and post-treatment evaluation remain undefined.[6]
Longevity biology: promising, unreplicated
The longevity case still rests substantially on a single 2020 prospective trial: 35 adults aged 64+ completing 60 sessions showed telomere elongation exceeding 20% across T-helper, cytotoxic T, NK and B cells (up to 37.6% in B cells) and a 37.3% reduction in senescent T-helper cells.[7] The biology is striking and the result is widely cited — but it has not been independently replicated at scale, there was no control arm, and no trial has yet linked those biomarker shifts to a clinical endpoint. Present it to patients as mechanistic and preliminary, not as a demonstrated anti-aging effect.
Protocol convergence and safety
| Indication | Sessions | Pressure / duration | Evidence grade |
|---|---|---|---|
| Persistent post-TBI symptoms | 40 | 90 min · ~2.0 ATA | Sham-controlled RCT, small n[1] |
| Chronic post-concussion cognition | 40–60 | 90 min · 2.0 ATA + air breaks | Retrospective cohort[2] |
| Long COVID | ~40 (median) | Standard multiplace protocol | Registry + mixed RCTs[3][6] |
| Cognitive enhancement (no injury) | — | — | Negative in largest RCT[5] |
| Longevity biomarkers | 60 | Daily, 5×/week | Preliminary, uncontrolled[7] |
On safety, the long-COVID registry is the most useful recent real-world dataset: 24 unsolicited adverse events across 232 treated patients, most commonly middle-ear barotrauma (11 patients). Solicited effects during treatment were fatigue (65%) and blurry vision (27%) — both reversible by three-month follow-up in all but one patient, who was found to have an incipient cataract. Ten patients discontinued because HBOT worsened their fatigue, which is a clinically meaningful signal in post-exertional-malaise phenotypes and worth screening for.[3]
Standard contraindications remain: untreated pneumothorax, certain chemotherapeutic agents, and uncontrolled conditions affecting middle-ear equalisation. Myopic shift and cataract progression should be part of informed consent for extended courses.
- Best supported: persistent symptoms after brain injury, at 40+ sessions.
- Reasonable with caveats: long COVID with cognitive and fatigue phenotypes, acknowledging registry-grade evidence and a real non-responder rate.
- Not supported by the best current trial: cognitive enhancement in uninjured or metabolically compromised older adults.
- Preliminary: longevity and senescence biomarkers — mechanistically interesting, clinically unproven.
- Dose matters more than pressure: under-dosed courses read as null across the literature.
Peak Human runs HBOT as a clinical protocol with defined endpoints — not a wellness add-on. Sit down with our team and map it against your indication.
Book Your HBOT ConsultationMedical note. This article is a summary of published research for educational purposes. It is not medical advice and does not establish a clinician–patient relationship. Hyperbaric oxygen therapy has specific contraindications and is not appropriate for everyone. Several of the findings above are preliminary, uncontrolled, or unreplicated, and one large controlled trial reported no benefit. Discuss any treatment decision with a qualified healthcare provider who knows your history.
References
- [1]Weaver LK, Ziemnik R, Deru K, et al. A double-blind randomized trial of hyperbaric oxygen for persistent symptoms after brain injury. Scientific Reports. 2025;15:6885. https://pubmed.ncbi.nlm.nih.gov/40011516/
- [2]Shabi Shlifer A, Suzin G, Shorer R, Lang E, Finci S, Elman-Shina K, Doenyas-Barak K, Efrati S. Hyperbaric oxygen therapy improves post-concussion symptoms in adults with childhood traumatic brain injury: a retrospective cohort study. Frontiers in Neurology. 2025;16:1641033. https://doi.org/10.3389/fneur.2025.1641033
- [3]Lansdorp CA, van Berkel J, Lalieu RC, et al. Hyperbaric oxygen therapy for long COVID: a prospective registry. Scientific Reports. 2025;15:28351. https://doi.org/10.1038/s41598-025-11539-0
- [4]Betyna-Białek M, Grobelska K, Borkowska A. The impact of hyperbaric therapy on cognitive functions in patients treated for idiopathic sudden hearing loss or hard-to-heal chronic wounds. Scientific Reports. 2026;16:1564. https://doi.org/10.1038/s41598-025-29302-w
- [5]Benari O, et al. Results of a double-blind sham-controlled trial examining the effects of hyperbaric oxygen therapy on cognition and brain biomarkers. Alzheimer's & Dementia. 2025. https://doi.org/10.1002/alz70861_108814
- [6]Minni A, et al. Hyperbaric Oxygen Therapy on Long COVID Symptoms: A Breath of Fresh Air. Diseases. 2026;14(2):60. https://doi.org/10.3390/diseases14020060
- [7]Hachmo Y, Hadanny A, Abu Hamed R, et al. Hyperbaric oxygen therapy increases telomere length and decreases immunosenescence in isolated blood cells: a prospective trial. Aging (Albany NY). 2020;12(22):22445–22456. https://pubmed.ncbi.nlm.nih.gov/33206062/