Ozone therapy

What it is

Medical ozone is produced in a generator from pure oxygen; the mixture contains only a few percent ozone. The best-known form is major autohemotherapy: blood is drawn from a vein, mixed with the gas and returned as an infusion. In minor autohemotherapy, a small amount of blood is injected into the muscle.

In addition, there are local applications. Ozone is injected into joints, muscles or intervertebral discs, used in dentistry as a gas, gel or oil, or applied to chronic wounds via bags. Some practices also offer rectal administration. These forms differ greatly, and the evidence has to be considered separately for each one.

How it is supposed to work

The explanation offered by proponents is called hormesis. Ozone reacts immediately with components of the blood and forms short-lived oxidation products. This small oxidative stress is supposed to reset the body’s own antioxidant systems, inflammatory signals and the oxygen release of red blood cells. The principle is known; fasting, cold or exercise work in a similar way as a dosed stimulus.

So far this has been shown mainly in cell culture and animal models. That the same chain runs in humans to a measurable extent and leads to freedom from symptoms has not been shown. For local injections, simpler explanations come into play: the gas has an antimicrobial effect, and in the intervertebral disc it is supposed to shrink tissue and reduce pressure on the nerve root.

What it is offered for

In Germany, ozone is offered mainly in private and naturopathic practices: for back and joint pain, chronic wounds, susceptibility to infection, exhaustion, long COVID and as general revitalization. For the broad promises such as more energy or a stronger immune system there are no controlled studies. For local pain treatments, however, there are, and they are the part that makes the discussion serious.

You can recognize reputable offers by a clear diagnosis before treatment, a written protocol with the ozone concentration, a test for G6PD deficiency before autohemotherapy, and strict hygiene. Promises that ozone cures infections, cancer or long COVID are not reputable.

What the studies show

Umbrella review 2026 — all meta-analyses

The authors searched for all meta-analyses of randomized studies comparing ozone with placebo, sham treatment or standard therapy. Of 1,243 hits, 7 meta-analyses on 4 applications remained. For diabetic foot, ozone was not superior for healing, relative risk 1.69 with a confidence interval of 0.90 to 3.17. For COVID-19, only the proportion of positive PCR tests decreased, not hospital admissions or mortality. Conclusion: the certainty of evidence was low or very low for all endpoints.

Cochrane 2015 — diabetic foot

3 randomized studies with 212 participants were included; the risk of bias was high or unclear. Compared with standard treatment, there was no detectable difference in healing, wound area, side effects or amputations. The authors felt unable to draw firm conclusions.

Knee meta-analysis 2026 — ozone versus cortisone

7 controlled studies with 409 people with knee osteoarthritis. The comparison was with an active treatment recommended in guidelines. Ozone reduced pain more strongly in the short and medium term. Because of the limited study quality, the authors did not want to make a definitive statement.

Where the data stop

Systemic use, that is, major autohemotherapy, is the least studied part, even though it is the most frequently sold. For long and post-COVID, the IGeL-Monitor found not a single study in 2023 and rated the service as unclear. For energy, immune defense, exhaustion or anti-aging there are no controlled data.

The positive findings also have limits. Many studies compared ozone with cortisone or with nothing at all, rarely with a sham injection. For herniated discs, the often-cited improvement of 4.25 points on the pain scale comes from uncontrolled before-and-after studies. Concentrations, amounts and number of sessions are hardly comparable. And side effects were reported so inconsistently in the meta-analyses that the authors could not say anything firm about them.

Status, approval and legal

In Germany, ozone therapy is not covered by statutory health insurance; you pay yourself. For major autohemotherapy, the IGeL-Monitor cites about 700 euros for approximately 10 treatments. It is offered by physicians and by Heilpraktiker (non-medical practitioners). There is no uniform regulation for concentration and procedure; the practitioners’ professional societies have their own protocols. In the US, the drug regulator FDA classifies ozone as a toxic gas with no known useful medical application.

Safety

When treatment is carried out correctly, serious incidents are rare according to current knowledge. A 2026 overview, however, lists a range of documented complications of autohemotherapy: breakdown of red blood cells and kidney failure at excessive concentrations, potassium imbalances, heart attacks, allergic reactions and infections due to hygiene errors. It recommends concentrations of 10 to 40 micrograms per milliliter and a test for G6PD deficiency beforehand. Gas entering an artery is dangerous. Cerebral gas embolisms with strokes have been described, for example in a 58-year-old patient during a disc injection, favored by a patent foramen ovale. Inflammation of the vertebrae has also been described after disc injections. Ozone must never be inhaled; it irritates the lungs. How often incidents occur nobody knows, because there is no registry.

Sources

• Cacciatore et al., Med Sci (Basel) 2026 (umbrella review)
• Liu et al., Cochrane Database Syst Rev 2015 (diabetic foot)
• Arias-Vázquez et al., Med Gas Res 2026 (knee osteoarthritis)
• de Sire et al., J Clin Med 2026 (shoulder)
• Kumar et al., Maedica (Bucur) 2024 (intervertebral disc)
• Casale et al., Clin Exp Rheumatol 2026 (safety of autohemotherapy)
• Khosravi, Mirzaasgari, BMJ Neurol Open 2024 (gas embolism)
• IGeL-Monitor 2023, ozone autohemotherapy for long/post-COVID
• FDA, 21 CFR 801.415