Molecular hydrogen · graded by use, eight uses
Molecular Hydrogen (Inhaled)
Graded on: condition-specific endpoints The outcome a trial measures to decide whether the treatment worked, such as heart attacks, blood pressure or a symptom score., one small trial each: COPD exacerbation symptom score, NAFLD liver fat, sleep scales.
Grade C: human trials exist, but they are small, mixed or not yet repeated. This is not a finding that it does not work. How to read a grade →
Compare this record with another, side by side →
How this letter has changedLatest: October 2026 · Changed.
One C covered COPD symptoms, fatty liver and sleep as if one finding, and the record omitted the only hard-endpoint trial. Each indication now has its own row at C as a single unreplicated small trial (Zheng et al., 2021; PMID PubMed ID: the permanent number the US National Library of Medicine gives a published paper. Each one links to that paper's record on PubMed. 33985501; Tao et al., 2022; PMID 35734974; Gao et al., 2026; PMID 40826930), with the sleep trial's funding by a device company, which took part in analysis and writing, stated beside the letter. HYBRID II (Tamura et al., EClinicalMedicine, 2023; PMID 36969346) missed its primary neurological outcome and is D on Route 2 for that use, with its favorable secondary outcomes and early stop stated. Ohsawa 2007 (PMID 17486089) is relabeled as cell culture and rat work. The endpoint type is corrected: the COPD and sleep scores are self-reported or actigraphic, only liver fat is a surrogate. Shi 2023 (PMID 37970554) and Ono 2017 (PMID 28669654) are added as context. The headline stays C. Applied the written grade definition under the editor's standing instruction of October 2026 (Arnold Sandlow, D.C.).PMID 17486089 PMID 28669654 PMID 33985501 PMID 35734974 PMID 36969346 PMID 37970554 PMID 40826930
Every change to a letter on this site is dated and published. The full evidence log →
This means breathing in hydrogen gas, often mixed with oxygen. Small single studies found benefits in several settings, such as lung disease flare-ups, fatty liver and sleep. None has been repeated. After cardiac arrest it did not improve recovery. In Parkinson's disease it did not help movement symptoms.
The C is shared by five of the eight uses graded on this record, each on its own evidence. The other three carry their own letters.
As studied, and the trials
Zheng et al. (Respir Res, 2021; PMID 33985501): double-blind Neither the participants nor the people assessing them know who received the real treatment until the trial ends., hydrogen and oxygen against oxygen alone; the self-reported symptom score fell 5.3 against 2.4 points, cough score also lower, lung function and blood gases unchanged. One trial, not replicated The same finding reproduced by a separate research group. A result seen once can be chance or bias; a replicated one is harder to explain away., self-reported endpoint, comparator oxygen alone, registered after it began: C.
As studied, and the trials
Tao et al. (J Cell Mol Med, 2022; PMID 35734974): 43 subjects, placebo-controlled; serum lipids and liver enzymes improved and liver fat on ultrasound and CT fell in moderate to severe cases. A single small trial on a surrogate: C.
As studied, and the trials
Gao et al. (Med Gas Res, 2026; PMID 40826930): single-blind, 7 days, actigraphic sleep time and efficiency improved and wake time fell against control, with better sleep and depression questionnaire scores. A device company funded the trial and took part in analysis and writing, stated beside the letter. One short single-blind trial: C.
As studied, and the trials
Tamura et al. (EClinicalMedicine, 2023; PMID 36969346), HYBRID II: the only inhaled hydrogen trial with a hard primary endpoint The one outcome a trial names in advance as its main test. Results on other outcomes carry less weight.. Good neurological function at 90 days, 56 against 39 percent, was not significant; 90-day survival (85 against 61 percent) and the modified Rankin scale favored hydrogen as secondary outcomes in a trial stopped early for pandemic enrollment restrictions. Funded by the gas supplier. Tested on its primary endpoint without a demonstrated benefit: D on Route 2, with the secondary signals stated. A completed trial at the planned size could move this.
As studied, and the trials
Lin et al. (J Pers Med, 2023; PMID 36675728): delirium on the Confusion Assessment Method in 10 of 83 patients who inhaled hydrogen (12 percent) against 17 of 70 controls (24 percent); C-reactive protein lower; pain scores slightly higher on day 1; length of stay and sleep unchanged. The abstract does not say whether the assessors were blinded, and 31 of the 184 randomized were not analyzed. One trial, not replicated: C.
As studied, and the trials
Wu et al. (2024; PMID 39099781): single-center randomized trial; brain edema volume on imaging before discharge was smaller and resolved faster with hydrogen than with oxygen alone, and sleep and pain scores were better on day 3. Imaging volume is a surrogate, from one unreplicated trial: C.
As studied, and the trials
Yoritaka et al. (Neurol Sci, 2021; PMID 34319514): double-blind pilot, hydrogen at 6.5 percent by volume (above the 4 percent flammability threshold cited elsewhere on this page) against placebo An inactive substitute made to look like the real treatment, so participants cannot tell which one they received. air for 16 weeks; no difference in the MDS-UPDRS total score. Tested on its primary endpoint without benefit: D on Route 2, low certainty.
As studied, and the trials
Chen et al. (2026; PMID 41908285): double-blind trial in 53 infants; adding hydrogen and oxygen nebulization to standard treatment did not improve spasm relief, EEG findings or IL-6. Elevated myocardial enzymes (28.3 against 10.9 percent), gastrointestinal dysfunction (43.5 against 23.9 percent) and abnormal liver function (8.7 against 0 percent) were more frequent with hydrogen. Tested on its primary endpoint without benefit: D on Route 2, with the adverse findings stated.
Read the grade that matches your goal. How EBL grades by use →
A letter describes the human evidence, not the product, the manufacturer, or the intervention's worth.
See the full evidence record →Funding: The single trial this Grade C for Sleep time and efficiency in people with sleep disorders rests on was funded by the maker of the device it tested, Dalian. The funder's role in design and analysis is not stated here, because the full trial report could not be read. The letter is what the trial supports under the grading rules; the funding is stated here so it can be weighed alongside it.
Funding: The single trial this Grade D for Good neurological outcome after out-of-hospital cardiac arrest rests on was funded by the maker of the device it tested. The funder's role in design and analysis is not stated here, because the full trial report could not be read. The letter is what the trial supports under the grading rules; the funding is stated here so it can be weighed alongside it.
Graded on Condition-specific endpoints, one small trial each: COPD exacerbation symptom score, NAFLD liver fat, sleep scales.
Why this gradeGrade C, small or unreplicated trials. The reasoning, and what the grade does not cover.
Grade C, one small randomized trial per indication and none replicated. Hydrogen and oxygen inhalation improved a self-reported symptom score in acute COPD exacerbation against oxygen alone (Zheng et al., 2021; PMID 33985501), liver fat in 43 people with fatty liver disease (Tao et al., 2022; PMID 35734974) and sleep measures over 7 days in a single-blind trial paid for by a device company (Gao et al., 2026; PMID 40826930). The one trial with a hard primary endpoint, HYBRID II after cardiac arrest, missed it in a study stopped early (Tamura et al., 2023; PMID 36969346), graded D on Route 2 on its own row with its favorable secondary outcomes stated. The selective-antioxidant mechanism (Ohsawa et al., 2007; PMID 17486089) is cell and rat work and carries no letter.
WHAT THIS GRADE DOES NOT COVERhydrogen-rich water, graded separately; any longevity endpoint; healthy adults, whom none of the graded trials enrolled; and any condition beyond the ones named, each tested once. Two further single trials are graded C as separate uses: fewer cases of delirium after noncardiac surgery in older adults (Lin et al., 2023; PMID 36675728) and less brain swelling on imaging after glioma surgery (Wu et al., 2024; PMID 39099781). Two trials found no benefit and are graded D on Route 2: motor symptoms in a 20-person Parkinson's pilot (Yoritaka et al., 2021; PMID 34319514) and spasm control in 53 infants with epileptic spasms, where heart, liver and digestive side effects were more frequent with hydrogen (Chen et al., 2026; PMID 41908285).
A grade describes the human evidence for a named use. It is not advice, and it is not a judgment on any individual prescription. How the letters are decided →
Each use, with its trials
This record carries a separate letter for each thing it has been measured on. The headline grade is the best of them, and it does not transfer to the others.
Zheng et al. (Respir Res, 2021; PMID 33985501): double-blind, hydrogen and oxygen against oxygen alone; the self-reported symptom score fell 5.3 against 2.4 points, cough score also lower, lung function and blood gases unchanged. One trial, not replicated, self-reported endpoint, comparator oxygen alone, registered after it began: C.
Tao et al. (J Cell Mol Med, 2022; PMID 35734974): 43 subjects, placebo-controlled; serum lipids and liver enzymes improved and liver fat on ultrasound and CT fell in moderate to severe cases. A single small trial on a surrogate: C.
Gao et al. (Med Gas Res, 2026; PMID 40826930): single-blind, 7 days, actigraphic sleep time and efficiency improved and wake time fell against control, with better sleep and depression questionnaire scores. A device company funded the trial and took part in analysis and writing, stated beside the letter. One short single-blind trial: C.
Tamura et al. (EClinicalMedicine, 2023; PMID 36969346), HYBRID II: the only inhaled hydrogen trial with a hard primary endpoint. Good neurological function at 90 days, 56 against 39 percent, was not significant; 90-day survival (85 against 61 percent) and the modified Rankin scale favored hydrogen as secondary outcomes in a trial stopped early for pandemic enrollment restrictions. Funded by the gas supplier. Tested on its primary endpoint without a demonstrated benefit: D on Route 2, with the secondary signals stated. A completed trial at the planned size could move this.
Lin et al. (J Pers Med, 2023; PMID 36675728): delirium on the Confusion Assessment Method in 10 of 83 patients who inhaled hydrogen (12 percent) against 17 of 70 controls (24 percent); C-reactive protein lower; pain scores slightly higher on day 1; length of stay and sleep unchanged. The abstract does not say whether the assessors were blinded, and 31 of the 184 randomized were not analyzed. One trial, not replicated: C.
Wu et al. (2024; PMID 39099781): single-center randomized trial; brain edema volume on imaging before discharge was smaller and resolved faster with hydrogen than with oxygen alone, and sleep and pain scores were better on day 3. Imaging volume is a surrogate, from one unreplicated trial: C.
Yoritaka et al. (Neurol Sci, 2021; PMID 34319514): double-blind pilot, hydrogen at 6.5 percent by volume (above the 4 percent flammability threshold cited elsewhere on this page) against placebo air for 16 weeks; no difference in the MDS-UPDRS total score. Tested on its primary endpoint without benefit: D on Route 2, low certainty.
Chen et al. (2026; PMID 41908285): double-blind trial in 53 infants; adding hydrogen and oxygen nebulization to standard treatment did not improve spasm relief, EEG findings or IL-6. Elevated myocardial enzymes (28.3 against 10.9 percent), gastrointestinal dysfunction (43.5 against 23.9 percent) and abnormal liver function (8.7 against 0 percent) were more frequent with hydrogen. Tested on its primary endpoint without benefit: D on Route 2, with the adverse findings stated.
The evidence, in fullEleven papers, each linked to its PubMed record.
Inhaled molecular hydrogen (usually a hydrogen and oxygen gas mixture) has a scattered human randomized trial base, one small trial per indication. Zheng et al. (Respiratory Research, 2021; PMID 33985501): multicenter double-blind trial in acute exacerbation of COPD, hydrogen and oxygen against oxygen alone; the Breathlessness, Cough and Sputum Scale, a self-reported symptom score, fell 5.3 against 2.4 points at day 7, with lung function and blood gases unchanged; the comparator was oxygen alone and the trial was registered after it began. Tao et al. (Journal of Cellular and Molecular Medicine, 2022; PMID 35734974): 43 subjects with NAFLD, 13 weeks; liver fat on imaging improved in moderate to severe cases, a surrogate. Gao et al. (Medical Gas Research, 2026; PMID 40826930): 66 people with sleep disorders, single-blind, 7 days; sleep time, efficiency and questionnaire scores improved; funded by the device company Jinkai Instrument (Dalian), which took part in data analysis, interpretation and writing. Shi et al. (J Clin Biochem Nutr, 2023; PMID 37970554): 64 patients with Omicron COVID-19, shorter viral shedding, one small trial. Ono et al. (J Stroke Cerebrovasc Dis, 2017; PMID 28669654): 50 patients with acute cerebral infarction, 3 percent hydrogen twice daily for 7 days against conventional intravenous medication, not placebo and not blinded; stroke scale and imaging improved. Tamura et al. (EClinicalMedicine, 2023; PMID 36969346), HYBRID II: 15-hospital double-blind placebo-controlled trial of 2 percent hydrogen for 18 hours after out-of-hospital cardiac arrest, stopped early at 73 patients; the primary outcome, good neurological function at 90 days, was 56 against 39 percent and not significant (relative risk The risk in one group expressed as a share of the risk in another. It can make a small absolute difference look large. 0.72, 95% CI 0.46 to 1.13, p=0.15); 90-day survival and the modified Rankin scale favored hydrogen as secondary outcomes; funded by the gas supplier Taiyo Nippon Sanso. Ohsawa et al. (Nature Medicine, 2007; PMID 17486089): cell culture and a rat ischemia model, the selective-antioxidant mechanism; preclinical, no weight for a human grade. Lin et al. (J Pers Med, 2023; PMID 36675728): 184 adults aged 65 and over having noncardiac surgery; delirium in 12 against 24 percent. Wu et al. (2024; PMID 39099781): single-center trial after glioma surgery; smaller brain edema volume on imaging. Yoritaka et al. (Neurol Sci, 2021; PMID 34319514): 20-person double-blind Parkinson's pilot, no difference on the MDS-UPDRS. Chen et al. (2026; PMID 41908285): 53 infants with epileptic spasms, no difference in spasm relief, more adverse findings with hydrogen. Grade C: single unreplicated small positives per indication, and the one hard-endpoint trial missed its primary outcome. Inhaled hydrogen is graded separately from hydrogen water (also C), a different delivery, dose and evidence base.
For the graded conditions, the measure the trial used: a symptom score in COPD exacerbation, liver fat imaging in fatty liver disease, actigraphic sleep time. Inhaled hydrogen concentration and session duration at every use: 1 to 4 percent is the range in the published trials; above 4 percent is flammable and has not been studied in humans. Oxidative stress markers are mechanism readouts, not graded endpoints.
Evidence notes: randomized trial, not independently replicated
Endpoint graded: Condition-specific endpoints, one small trial each: COPD exacerbation symptom score, NAFLD liver fat, sleep scales (self_report)
Last reviewed: July 2026
Educational use only · Not medical advice · EBL grades published human evidence and does not recommend treatment. Stopping a prescription is a decision for your physician.
Primary evidenceEleven papers, each linked to its PubMed record.
Inhaled molecular hydrogen (usually a hydrogen and oxygen gas mixture) has a scattered human randomized trial base, one small trial per indication. Zheng et al. (Respiratory Research, 2021; PMID 33985501): multicenter double-blind trial in acute exacerbation of COPD, hydrogen and oxygen against oxygen alone; the Breathlessness, Cough and Sputum Scale, a self-reported symptom score, fell 5.3 against 2.4 points at day 7, with lung function and blood gases unchanged; the comparator was oxygen alone and the trial was registered after it began. Tao et al. (Journal of Cellular and Molecular Medicine, 2022; PMID 35734974): 43 subjects with NAFLD, 13 weeks; liver fat on imaging improved in moderate to severe cases, a surrogate. Gao et al. (Medical Gas Research, 2026; PMID 40826930): 66 people with sleep disorders, single-blind, 7 days; sleep time, efficiency and questionnaire scores improved; funded by the device company Jinkai Instrument (Dalian), which took part in data analysis, interpretation and writing. Shi et al. (J Clin Biochem Nutr, 2023; PMID 37970554): 64 patients with Omicron COVID-19, shorter viral shedding, one small trial. Ono et al. (J Stroke Cerebrovasc Dis, 2017; PMID 28669654): 50 patients with acute cerebral infarction, 3 percent hydrogen twice daily for 7 days against conventional intravenous medication, not placebo and not blinded; stroke scale and imaging improved. Tamura et al. (EClinicalMedicine, 2023; PMID 36969346), HYBRID II: 15-hospital double-blind placebo-controlled trial of 2 percent hydrogen for 18 hours after out-of-hospital cardiac arrest, stopped early at 73 patients; the primary outcome, good neurological function at 90 days, was 56 against 39 percent and not significant (relative risk 0.72, 95% CI 0.46 to 1.13, p=0.15); 90-day survival and the modified Rankin scale favored hydrogen as secondary outcomes; funded by the gas supplier Taiyo Nippon Sanso. Ohsawa et al. (Nature Medicine, 2007; PMID 17486089): cell culture and a rat ischemia model, the selective-antioxidant mechanism; preclinical, no weight for a human grade. Lin et al. (J Pers Med, 2023; PMID 36675728): 184 adults aged 65 and over having noncardiac surgery; delirium in 12 against 24 percent. Wu et al. (2024; PMID 39099781): single-center trial after glioma surgery; smaller brain edema volume on imaging. Yoritaka et al. (Neurol Sci, 2021; PMID 34319514): 20-person double-blind Parkinson's pilot, no difference on the MDS-UPDRS. Chen et al. (2026; PMID 41908285): 53 infants with epileptic spasms, no difference in spasm relief, more adverse findings with hydrogen. Grade C: single unreplicated small positives per indication, and the one hard-endpoint trial missed its primary outcome. Inhaled hydrogen is graded separately from hydrogen water (also C), a different delivery, dose and evidence base.
Products carry no letter. The letter belongs to the therapy as the trials studied it, and no trial tested these products unless the entry says so. Listed alphabetically; payment does not affect order.
High-output H₂ inhalation generator. US-based company with good customer support and independently measured output specifications. For dedicated H₂ inhalation protocols.
This page carries a letter. EBL takes no commission, referral fee or payment of any kind on anything it grades. No letter has been bought and none is for sale. Products listed under this modality may be commercial partners, and their purchase links may earn EBL a commission. Those products carry no letter of their own, every relationship is disclosed beside the link, and where a product does not deliver what the trials tested, this page says so regardless of who pays.
How this is enforced
That is not a promise, it is a build condition. Two checks run on every deploy: one refuses to publish if a commission is claimed on anything graded, the other refuses if a page names a commercial relationship that is not declared in the register. If either fails, this page does not go live. You can read how the checks work, and what they do not cover, on the evidence integrity page.
Evidence Based Longevity, Inc. is owner-operated and edited by Arnold Sandlow, D.C., a California-licensed chiropractor. Evidence Based Longevity publishes in English only · Machine translations are not produced or reviewed by EBL for accuracy
© Evidence Based Longevity, Inc.. All rights reserved
These statements have not been evaluated by the Food and Drug Administration. Any grade on this page describes the strength of published human trial evidence in the populations studied. It is not a recommendation for any individual and is not medical advice. Starting, stopping or changing any supplement or medication is a decision for your physician.
How EBL searched for the trials (1 search)
The search that was run, what it returned, and what was left out and why, so the absence of evidence can be checked as well as its presence.
(hydrogen[tiab]) AND (inhal*[tiab]) AND (randomized controlled trial[pt] OR randomized[tiab] OR randomized[tiab]) AND humans[mh]Grades move when new trials publish. A free monthly email lists every grade EBL moved, with the reason and the trial behind it: the sign-up is on the Evidence Log →