Longevity Science · Antioxidants · Hydrogen Therapy
Most antioxidants don't discriminate — they scavenge beneficial signalling molecules alongside harmful ones. Molecular hydrogen does something different: it preferentially neutralises the hydroxyl radical, the most destructive reactive oxygen species in your body, while leaving protective ROS intact. Here is what the science actually says.
In 2007, a team led by Shigeo Ohsumi and Ikuroh Ohsawa at Tokyo Metropolitan Institute for Medical Science published a landmark paper in Nature Medicine demonstrating that inhaled hydrogen gas (1–4%) dramatically reduced ischaemia-reperfusion brain injury in rats. The mechanism they proposed — and which has since been extensively studied — rests on a key chemical principle: selective reactivity.
Not all ROS are equal. Your cells generate several species continuously:
Broad antioxidants like vitamin C and NAC intercept multiple species including H₂O₂, which may paradoxically blunt beneficial signalling (the "antioxidant paradox" discussed extensively in exercise physiology literature — see Ristow et al., PNAS, 2009). Molecular hydrogen reacts readily with •OH and ONOO⁻ but is essentially inert to O₂•⁻ and H₂O₂ under physiological conditions. This selectivity is its defining property.
Subsequent research has proposed mechanisms beyond direct radical scavenging, though these remain less settled:
Honest caveat: most mechanistic data comes from cell culture and rodents. Extrapolating mechanisms to humans requires caution.
Human trials of molecular hydrogen are mostly small, short-duration, and from a handful of Japanese research groups. This limits generalisability. Here is the honest picture by condition.
Nakao et al. (2010, Nutrition Research) conducted one of the earliest randomised placebo-controlled crossover trials in 20 patients with potential metabolic syndrome risk factors. Subjects drank 1.5–2L of hydrogen-rich water (HRW, ~1.6 mg/L) daily for 8 weeks. Results: significant reduction in urinary 8-isoprostane (marker of lipid oxidation), trend toward improved HDL/LDL ratio. This was a small crossover trial — not definitive, but mechanistically coherent.
Kajiyama et al. (2008, Nutrition Research) tested HRW in 30 type 2 diabetics: hydrogen water improved glucose tolerance and reduced HbA1c relative to control, with improved urinary oxidative stress markers. Again, small sample, but consistent with the antioxidant mechanism.
Yoritaka et al. (2013, Movement Disorders) ran a randomised, double-blind, placebo-controlled pilot in 17 patients with Parkinson's disease: HRW (1L/day, ~1 mg/L) vs. plain water for 48 weeks. The HRW group showed no worsening on UPDRS motor scores vs. the placebo group which deteriorated — a notable finding given the progressive nature of PD. The trial was explicitly a pilot to guide future work, powered for feasibility not effect size. A follow-up open-label phase saw some regression. The signal is interesting; the evidence is preliminary.
Aoki et al. (2012, Medical Gas Research) tested HRW in elite soccer players: decreased blood lactate post-exercise and reduced muscle fatigue markers. Ostojic et al. (2011, Journal of Sports Medicine and Physical Fitness) found reduced lactate after maximal exercise with HRW. However, subsequent blinded trials have not consistently replicated performance gains. A 2020 meta-analysis (LeBaron et al.) found modest effects across studies but noted high heterogeneity and publication bias risk.
Two small Japanese trials in cancer patients undergoing radiotherapy found that HRW consumption reduced markers of oxidative stress and quality-of-life scores improved (Kang et al., 2011; Zhao et al., 2014). This is a context where reducing radiation-induced •OH damage is mechanistically logical. Evidence quality: moderate for the surrogate marker outcomes.
Many animal studies show impressive H₂ effects in stroke, cardiovascular disease, liver injury, and ageing models. Until human RCTs replicate these findings, caution is warranted. Marketing claims citing rodent data as if they apply to humans are common and misleading.
| Condition | Key Study | H₂ Dose / Form | Primary Outcome | Evidence Quality |
|---|---|---|---|---|
| Brain ischaemia (rat) | Ohsawa et al., Nature Med. 2007 | 2% inhaled H₂, 2h | ↓ infarct size, ↓ oxidative stress markers | Strong (animal) |
| Metabolic syndrome / oxidative stress | Nakao et al., Nutr. Res. 2010 | HRW 1.5–2L/day, ~1.6 mg/L, 8 wk | ↓ 8-isoprostane, trend ↑ HDL | Moderate (small RCT) |
| Type 2 diabetes | Kajiyama et al., Nutr. Res. 2008 | HRW 900 mL/day, 8 wk | ↓ HbA1c, ↓ oxidative stress | Moderate (small RCT) |
| Parkinson's disease | Yoritaka et al., Mov. Disord. 2013 | HRW 1L/day, ~1 mg/L, 48 wk | No worsening UPDRS vs. control decline | Pilot (n=17) |
| Athletic performance / lactate | Aoki et al., Med. Gas Res. 2012 | HRW 1.5L/day, 1 wk | ↓ blood lactate post-sprint | Weak (mixed replication) |
| Radiation-induced fatigue (cancer) | Kang et al., Med. Gas Res. 2011 | HRW 1.5–2L/day during RT | ↓ fatigue, improved QoL scores | Moderate (small RCT) |
| Non-alcoholic fatty liver (animal) | Nishimura et al., Hepatol. Res. 2018 | HRW, 8 wk (mouse) | ↓ liver oxidative damage, ↓ steatosis | Strong (animal) |
| Cardiovascular inflammation | Multiple rodent studies | Various | ↓ NF-κB, ↓ inflammatory cytokines | Weak (no human RCTs) |
HRW = hydrogen-rich water. RCT = randomised controlled trial. Evidence quality ratings reflect human trial robustness only — strong animal evidence does not imply human efficacy.
The form of H₂ delivery matters enormously — and this is where most consumer products diverge from the research dosing used in clinical trials.
Portable electrolysis units split water to produce dissolved H₂ at the point of use. Quality units reach 1.0–1.6 mg/L (1000–1600 ppb), at or near saturation at atmospheric pressure. This matches trial dosing. The key limitation is off-gassing: hydrogen escapes any unsealed container rapidly. Drink immediately after generation. Generator cost ranges from $80–$400 USD depending on quality.
Mg + 2H₂O → Mg(OH)₂ + H₂↑. Tablets dissolved in a sealed bottle can achieve 1.0–1.6 mg/L if the container is properly sealed and gas is allowed to build pressure. Cheaper per dose than generators. The reaction also elevates pH slightly (more alkaline water), which is physiologically irrelevant at these concentrations but is often marketed separately. Reputable tablet products specify H₂ concentration at dissolution; be sceptical of brands that do not.
The most consumer-accessible form is often the least reliable. Unless hydrogen water is stored in aluminium pouches under pressure (similar to IV bags) or special hydrogen-impermeable bottles, significant H₂ has off-gassed before you open it. Several lab analyses of retail hydrogen water have found concentrations near zero. If using pre-bottled products, choose aluminium pouch formats from transparent brands with published H₂ testing data.
Clinical trials for ischaemic conditions often use inhalation of 2–4% H₂ (mixed with oxygen), which delivers H₂ systemically via the lungs at much higher rates than drinking. This is the standard in hospital-grade intervention research. Consumer inhalation devices exist but are expensive ($500–$2000+) and not validated outside clinical settings. Below 4% H₂ in air is non-flammable. This is the delivery method with the strongest mechanistic data; it is also the least accessible.
Dissolved H₂ is measured in mg/L (equivalent to ppm) or in ppb (µg/L). The saturation point of H₂ in water at room temperature and atmospheric pressure is approximately 1.57 mg/L (1570 ppb). Most clinical trials used 0.8–1.6 mg/L. A reagent-drop test kit (methylene blue reduction) gives a rough qualitative check; electronic H₂ meters provide more precise readings. If you are spending money on hydrogen water products, a basic H₂ meter ($30–$60) is worth having to verify what you are actually consuming.
This is where hydrogen water marketing runs furthest ahead of the evidence. Here is what the data actually supports:
Several small trials (Aoki 2012; Ostojic 2011) found reduced post-exercise blood lactate in athletes consuming HRW. The mechanistic rationale: H₂ may reduce oxidative stress-induced impairment of mitochondrial function, allowing better lactate clearance. However, a 2018 blinded trial by Botek et al. found no significant effect on exercise performance or lactate in trained cyclists. The conflicting results likely reflect heterogeneous participant populations, H₂ concentrations, and blinding quality.
A 2014 trial by Kawamura et al. found reduced markers of muscle micro-damage and lower perceived soreness with HRW post-eccentric exercise. This is consistent with the antioxidant mechanism (exercise-induced ROS contributing to DOMS). Effect sizes were modest. This is among the more plausible athletic applications given the clear oxidative stress component of DOMS.
No well-powered trial has examined whether H₂ impairs training adaptations (the hormetic concern — whether reducing exercise-induced ROS blunts mitochondrial biogenesis). This is a legitimate open question. Short-term supplementation around individual sessions (acute dosing) is less likely to impair adaptations than chronic around-the-clock antioxidant supplementation, but the question has not been directly tested with H₂.
Bottom line for athletes: H₂ water around hard training sessions is low-risk and mechanistically plausible for recovery. Performance gains are not established by the current literature. Do not reduce training load in anticipation of H₂ doing the work.