Breathing pure oxygen inside a pressurised chamber sounds dramatic — and the claims made for it often are. From professional athletes recovering from torn ligaments to burn victims seeking accelerated wound closure, hyperbaric oxygen therapy (HBOT) has attracted interest across an unusually wide range of injury contexts. But what does the clinical literature actually support, and where does the evidence fall short?

What Is Hyperbaric Oxygen Therapy?

HBOT involves breathing 100% oxygen inside a chamber pressurised to greater than one atmosphere — typically 2.0 to 2.5 atmospheres absolute (ATA). Under these conditions, the partial pressure of oxygen in arterial blood rises from its normal range of 75–100 mmHg to between 1,200 and 2,000 mmHg. This dramatic elevation in dissolved plasma oxygen, achieved independent of haemoglobin, allows oxygen to diffuse into tissues that standard circulation cannot adequately reach — particularly hypoxic, damaged, or poorly perfused wound beds.

The therapy was first used clinically in the 1950s and now has 14 approved indications according to the Undersea and Hyperbaric Medical Society (UHMS), including decompression sickness, carbon monoxide poisoning, necrotising fasciitis, and wound healing in select contexts.

Standard session protocols involve 60–120 minutes at pressure, repeated daily or on alternating days, with courses typically ranging from 10 to 40 or more sessions depending on the indication.

Why the Mechanism Is Compelling

Tissue healing is fundamentally oxygen-dependent. Fibroblasts require O₂ to hydroxylate proline and lysine — the steps that generate stable Type I collagen. Angiogenesis — the formation of new blood vessels that restores perfusion to injured tissue — depends on vascular endothelial growth factor (VEGF), whose synthesis is stimulated by HBO. Neutrophil bactericidal activity drops sharply below a tissue pO₂ of approximately 30 mmHg, creating the hypoxic-infectious wound cycle common in burns and post-surgical complications.

HBOT addresses all of these simultaneously. It also triggers what researchers call the “hyperoxic–hypoxic paradox”: repeated cycling from high oxygen back to room air is sensed at the cellular level as a relative hypoxic signal, upregulating HIF-1α (hypoxia-inducible factor), which in turn activates repair genes including VEGF, erythropoietin, and angiopoietin-2. The result is a sustained pro-repair cellular state that persists between treatment sessions.

Four Mechanisms Behind HBOT’s Healing Effect

This mechanistic clarity has made HBOT appealing across multiple injury types. The challenge — as we explore below — is that a compelling mechanism does not automatically produce compelling clinical results in every application.

What the Evidence Shows

Evidence Overview

●●●● High  |  ●●●○ Moderate  |  ●●○○ Low  |  ●○○○ Very Low  —  ↑ Supportive  |  ↗ Trending positive  |  → Neutral

Burns: Historically Recognised, Evidentially Limited

Burn wound healing is one of HBOT’s longest-established applications, and UHMS officially recognises it as an approved indication. The central piece of clinical evidence is a small RCT from 1974 (Hart and colleagues) that found mean healing times of 19.7 days with HBOT versus 43.8 days in controls — a striking result.

However, the Cochrane review that consolidated this evidence found only two qualifying RCTs, both of poor methodological quality. A larger and more rigorous trial (Brannen 1997) found no significant difference in mortality, length of stay, or number of surgeries once patient condition was adjusted for. The Cochrane authors concluded that the available evidence is insufficient to provide clear practice guidelines and called for larger, better-designed trials.

HBOT for burns is clinically used and intuitively plausible, but the evidence has not kept pace with the clinical adoption. This does not mean HBOT does not work for burns — it means we cannot yet say with confidence that it does.

Post-Operative Recovery: Context Is Everything

The picture for post-operative use is more nuanced and more indication-specific than most HBOT advocates acknowledge.

For osteoradionecrosis (ORN) — a debilitating complication of head and neck radiation — the evidence is the strongest in the entire HBOT healing portfolio. A Cochrane review of 14 trials (753 participants) found that HBOT improved the likelihood of mucosal coverage (risk ratio 1.3; 95% CI 1.1–1.6; number needed to benefit = 5) and substantially reduced wound breakdown in operative cases. This represents moderate-quality evidence — not definitive, but meaningfully above the Low quality that characterises most of the field.

For compromised skin grafts and flaps, a systematic review of four RCTs found a striking improvement in split-skin graft survival in burns patients (RR 3.50), but all trials were rated at high or unclear risk of bias, and smaller trials in other graft/flap contexts showed no benefit. UHMS guidance is appropriately specific: HBOT is indicated for grafts and flaps that are compromised by vascular insufficiency, radiation, or hypoxia — not for uncompromised tissue where perfusion is adequate.

For general post-operative wound healing, a 2020 systematic review of 13 RCTs (627 patients) found that 10 showed benefit for at least one outcome, but the outcomes were highly varied and meta-analysis was impossible due to heterogeneity. This is a promising signal, not a conclusion.

Tendons and Ligaments: Strong Animal Evidence, Thin Human Data

This is where HBOT’s reputation — particularly in sports medicine — has grown faster than its evidence base.

A 2025 systematic review (PMID: 41036684) synthesising both animal and human studies found consistent positive effects of HBOT on tendon healing, including increased collagen density, fibre alignment, and synthesis. In a rabbit model of patellar tendinopathy, HBOT-treated tendons showed 34.8% greater ultimate tensile load and 82.2% higher hydroxyproline concentrations at 10 weeks. Rotator cuff models in rabbits showed improved vascularisation and tensile strength at the tendon-bone interface. These results are biologically coherent and methodologically consistent.

For ligaments specifically, a 2024 rabbit ACL reconstruction study (Leite et al.) found that HBOT-treated animals demonstrated improved graft maturation on MRI, reduced tunnel widening, and greater load to failure compared to controls. A 2007 study showed enhanced collagen fibre compaction and higher pullout strength. A comparative non-randomised clinical study in rugby players with grade 2 MCL injuries reported accelerated recovery, though the non-randomised design limits interpretation.

The problem is that these are animal data. Rabbit tendons and human tendons are not the same structure, healing timelines differ, and controlled laboratory conditions cannot replicate clinical complexity. No qualifying human RCT for HBOT in tendon or ligament healing has been published. The 2025 systematic review explicitly states that larger human trials and standardised protocols are needed before clinical recommendations can be made.

Fractures: No Clinical Evidence Exists

For bone fractures and non-union, the Cochrane review is unambiguous: zero qualifying human randomised controlled trials have been identified. The biological rationale exists — HBOT upregulates BMP-2 and VEGF, stimulates osteoblast differentiation, and promotes type H vessel formation in bone — but these findings are entirely preclinical. Three RCTs were registered as underway at the time of the 2012 Cochrane update; none appear to have published qualifying results since.

In the absence of any human trial data, HBOT cannot be considered a clinical option for fracture management outside of formally registered research protocols.

Who Is This Most Relevant For?

HBOT’s current evidence base suggests it is most clinically relevant for:

• Patients with osteoradionecrosis or late radiation tissue injury, where Cochrane evidence provides moderate-quality support.
• Patients with compromised flaps or grafts in the context of radiation damage, vascular compromise, or hypoxia, where expedient HBOT initiation may salvage threatened tissue.
• Patients with complex thermal burns at specialist centres where HBOT infrastructure exists and patient selection criteria can be applied.
• Athletes or rehabilitation patients considering HBOT for tendon or ligament recovery should understand that while the preclinical signal is consistent, there are no qualifying human RCTs — and that HBOT protocols for these indications are unstandardised.

The evidence is weakest — and in the case of fractures, absent — for patients considering HBOT primarily for musculoskeletal healing without the vascular compromise or radiation injury that drives the strongest evidence.

What the Research Does Not Yet Tell Us

The HBOT injury-healing evidence base has several meaningful gaps that limit clinical translation:

• Optimal protocols are undefined for all musculoskeletal indications. Pressures ranging from 1.5 to 3.0 ATA, session lengths from 60 to 120 minutes, and courses from 10 to 60 sessions have all been used. Without head-to-head protocol comparisons, clinicians cannot determine whether a 20-session versus 40-session course is meaningfully different.
• The fracture evidence deficit is complete. No human RCT data exist, and this has not changed since the 2012 Cochrane review.
• Tendon and ligament human RCTs are absent. The translation gap between rabbit models and human clinical benefit remains entirely uncrossed.
• Patient selection criteria for burns and post-surgical HBOT are not established. The Hart 1974 effect in burns has not been reproducible in other trials, suggesting there may be a subset of patients who benefit disproportionately — but that subset is unidentified.
• Long-term outcomes beyond the immediate healing period are rarely reported.

Practical Context

HBOT should be thought of as a specialist adjunct, not a standalone treatment. It is most defensible when standard care has limitations — when tissue is poorly perfused, when radiation has damaged the healing environment, or when a graft or flap is at risk of failure. In these contexts, it addresses a genuine physiological deficit.

For routine musculoskeletal injuries where perfusion is intact and healing is progressing normally, the case for HBOT is considerably weaker. The cost, logistics, and access requirements of repeated hyperbaric sessions are not trivial, and the evidence does not yet support HBOT as a standard adjunct to ACL reconstruction, Achilles tendon repair, or fracture management.