The Ohsawa 2007 Study, Explained

If you have ever read anything about molecular hydrogen therapy, you have read about the Ohsawa 2007 paper. It is the citation that shows up in every review, every clinical study, every marketing claim. It is, more than any other, the foundational document of the field.

But most summaries of the paper are thin. "Hydrogen is an antioxidant" is true, in the way that "the heart pumps blood" is true — it omits the interesting parts. This article walks through what the paper actually did, what it found, what it didn't show, and the field of research that grew from it.

Why this paper matters

Before 2007, hydrogen was an industrial gas. It was used in fuel cells, in chemical manufacturing, in welding. It was not, in any mainstream sense, a therapeutic molecule. The idea that inhaling a small amount of H₂ could treat disease was not on anyone's research agenda.

What Ohsawa and colleagues at Nippon Medical School proposed in their June 2007 paper was that hydrogen, despite its chemical simplicity, could act as a selective antioxidant in mammals — reducing the most cytotoxic reactive oxygen species (the hydroxyl radical) while leaving the signaling ROS alone. The proposal was not just new; it cut against the prevailing wisdom that any antioxidant intervention would be a blunt instrument. Ohsawa's hypothesis was that the smallest molecule in the universe could be the most precise one.

That single paper launched the field. The bibliometric record now includes more than a thousand peer-reviewed publications on molecular hydrogen, 80+ registered clinical trials, and reviews in every major physiology journal.^[3] None of that work would exist in its current form without the 2007 foundation.

The paper, in citation form

The research team

The first author, Ikuroh Ohsawa, was a researcher at the Department of Biochemistry and Cell Biology, Institute of Development and Aging Sciences, Nippon Medical School, Kawasaki, Japan. The senior author, Jun-ichi Ohta, led a group that had been working on oxidative stress and neurodegeneration for several years. The other authors — Masahiko Ishikawa, Kumiko Takahashi, Megumi Takahashi, Kazuhiro Nishimaki, Keiji Yamagata, Ryoichi Katsumura, and Yasuhiro Nishimaki — covered the biochemistry, the animal model, and the analytical chemistry that the study required.

It is worth noting what kind of institution this was. Nippon Medical School is a research-oriented medical university, and the Institute of Development and Aging Sciences is focused on the cellular biology of aging. The choice of H₂ as a candidate molecule came out of work on the oxidative stress theory of aging — the proposal that accumulated damage from reactive oxygen species is a primary driver of age-related decline. If H₂ could selectively reduce the most damaging ROS, the team reasoned, it might be doing something fundamentally new.

What they did

The study had three parts: a cell-culture experiment, a rat model of stroke, and a rat model of oxidative damage to the liver by a known toxin (anti-thyroid drug, carbon tetrachloride — used to induce oxidative stress).

The cell-culture experiment

The team exposed cultured cells to a system that generates hydroxyl radicals (·OH) — the most damaging of the reactive oxygen species — and asked whether dissolved H₂ would reduce the resulting damage. The hydroxyl radical generator used was based on a Fenton-like reaction (iron + hydrogen peroxide), which produces ·OH in known quantities.

When hydrogen was dissolved in the culture medium, the cells were protected. The hydroxyl radical signal dropped, and markers of oxidative damage (lipid peroxidation, protein carbonylation) were reduced.

The rat stroke model

The team then used a standard rat model of focal cerebral ischemia-reperfusion — the same kind of injury that happens in human stroke when blood flow is restored after being cut off. In this model, much of the cellular damage happens not during the ischemia itself but during the reperfusion, when a flood of reactive oxygen species is generated as oxygen returns to oxygen-starved tissue.

Rats were exposed to 2% or 4% hydrogen gas (mixed with air) for 30 minutes, beginning shortly after the ischemic event. The control group breathed ordinary air. The hydrogen-exposed group showed smaller infarct volumes (the volume of brain tissue that died), less oxidative damage, and better behavioral outcomes on standardized neurological tests.

The carbon tetrachloride model

As a second test of the general principle, the team used a rat model of carbon tetrachloride (CCl₄) liver injury, a classic oxidative-damage model. CCl₄ is metabolized in the liver to produce the trichloromethyl radical (·CCl₃) and other ROS, causing liver damage. Hydrogen exposure reduced the markers of liver injury in this model as well.

What they found

The headline finding was a single sentence in the abstract: "Hydrogen selectively reduced the hydroxyl radical, the most cytotoxic of reactive oxygen species (ROS), and effectively protected cells; however, H₂ did not react with other ROS, which possess physiological roles."^[1]

That sentence did three things at once. It proposed a mechanism. It bounded the mechanism — H₂ was not a general antioxidant. And it explained why that bounded mechanism was the interesting one: the other ROS weren't just bystanders, they were doing useful work, and any intervention that interfered with them risked interfering with normal cell signaling.

In the cell-culture work, the team also showed that hydrogen dissolved in culture medium was protective at concentrations compatible with what could be achieved in vivo through inhalation. The transition from a petri dish to a living rat brain was, mechanistically, plausible.

Why selectivity matters

To appreciate why the selectivity finding was important, it helps to know what the prevailing model of antioxidant therapy looked like in 2007.

Reactive oxygen species fall into a few categories. Hydrogen peroxide (H₂O₂) is a signaling molecule — your immune cells use it to coordinate attacks, your mitochondria use it to talk to the nucleus, your muscles use it to adapt to training. Superoxide (O₂⁻) and nitric oxide (·NO) are also signaling molecules. The hydroxyl radical (·OH), produced as a byproduct of Fenton chemistry, is the one that is purely destructive. It reacts indiscriminately with whatever it touches and there is no enzyme in the body that degrades it.

The antioxidant supplements that existed in 2007 — vitamin C, vitamin E, glutathione, N-acetylcysteine, resveratrol, and others — were all broad-spectrum. They reacted with multiple ROS, including the signaling ones. This is the technical reason why high-dose antioxidant supplementation has produced mixed and sometimes negative results in clinical trials: in some protocols, blunting the signaling ROS interferes with the adaptive response the body is trying to mount.

Hydrogen's selective profile — neutral toward H₂O₂, superoxide, and ·NO; reactive only toward ·OH — sidestepped this problem. The molecule could in principle reduce the most damaging ROS without interfering with the signaling ones. It was, in the language of the field, the first "selective antioxidant."

What the paper did not show

For a fair reading of the Ohsawa paper, it is worth being explicit about its limits. The study established a mechanism in cells and showed the mechanism produced measurable protection in two rat models. It did not establish that hydrogen would have the same effects in humans. It did not establish the dose. It did not establish the optimal route. It did not establish a safety profile for chronic exposure.

These are not criticisms of the paper. The 2007 study was a foundational mechanistic paper, and the 2007 team did exactly what a foundational mechanistic paper is supposed to do: identify a phenomenon, characterize it cleanly, and leave the clinical and translational work to follow. The follow-up work has now been ongoing for 18 years.

What it kicked off

The Ohsawa 2007 paper was the seed of a research program that has run for nearly two decades. A few of the major threads:

Human safety

The first non-trivial question after a preclinical result is whether the molecule is safe in humans. That work has now been done thoroughly. Cole et al. (2021) ran a Phase 1 study in which healthy adults were exposed to 2.4% H₂ gas for up to 72 hours with no clinically significant adverse events.^[2] Todorovic et al. (2023) aggregated 64 human studies and 81 clinical trials in a comprehensive review, concluding that hydrogen administration is safe across every delivery method tested.^[3]

Delivery methods

The original 2007 paper used inhaled gas. Subsequent work has explored drinking hydrogen-rich water, hydrogen-rich saline, hydrogen-producing tablets, hydrogen machines, and the stomach-reaction capsule approach. Each has its own dose-retention profile, and the practical question of which method delivers the most H₂ to the bloodstream is still being characterized.

Clinical indications

The clinical literature has grown to cover exercise recovery, metabolic health, hepatic function, neuroprotection, cardiovascular markers, and more. The Kajiyama 2008 study — one of the first human trials, in 60 patients with type 2 diabetes or impaired glucose tolerance — found significant improvements in lipid and glucose metabolism and reductions in TNF-α, IL-6, and CRP after eight weeks of hydrogen-rich water.^[7] The Botek 2022 and Sládečková 2024 studies extended this to exercise-induced oxidative stress and recovery in trained athletes.^[4][5]

Mechanism expansion

Subsequent work has proposed additional mechanisms beyond the selective antioxidant idea: gene expression modulation (particularly of Nrf2-mediated antioxidant response pathways), anti-apoptotic signaling, mitochondrial membrane protection, and modulation of inflammatory cascades. The selective antioxidant story is still the foundation, but the picture has filled in.

The current state of H₂ research

By 2023, the Todorovic et al. review of the field had this to say about the cumulative evidence: "All the studies have unequivocally confirmed the safety and legitimacy of H₂ in human consumption in all the administration methods tested, and therefore set the foundation for clinical trials for a variety of indications."^[3]

The honest summary, as of mid-2026: the safety record is strong, the mechanism is well-characterized at the preclinical level, and a growing number of human studies are showing reproducible signals in specific indications. The clinical evidence is not yet at the level of, say, an FDA-approved indication — but it is well past the level of speculation. Molecular hydrogen has earned its place in the wellness-research conversation.

What this means for a daily routine

The Ohsawa 2007 paper matters because it gave the field a mechanism worth building on. For someone deciding whether to add molecular hydrogen to their routine, the practical question is not whether the molecule does what the paper says it does — the paper's basic mechanism has been replicated and extended many times over. The practical question is delivery: which method gets a consistent dose of H₂ to your bloodstream, every day, at a fair price, without the loss-to-air problem that affects hydrogen water and tablets.

For our answer to that question, the hydrogen water vs. hydrogen capsules comparison walks through the engineering. For a broader look at the molecule and the field, the primer on molecular hydrogen therapy is the place to start.