What hydrogen water is, in one number: dissolved H2, measured in ppm
Hydrogen water is plain water with molecular hydrogen (H2) gas dissolved into it, usually by electrolysis.
The one number that makes any of it checkable is concentration: dissolved hydrogen in milligrams per liter, or parts per million. The Molecular Hydrogen Institute puts the unit question plainly: "In dilute concentrations, 1 ppm is about the same as 1 mg/L and they are often used interchangeably."
The same page gives a general figure for the field: "Studies using hydrogen gas dissolved in water range from 0.5 mg/L to 1.6+ mg/L, with most studies using a concentration near 1.6 mg/L," and describes 1.6 mg/L as "the concentration at saturation" in the scientific literature.
The trials cited on this page do not all sit inside that figure. The 2020 Scientific Reports trial measured its water at 0.753 mg/L, inside the range; the seven trials pooled for lipids in 2023 list concentrations up to 15 ppm, far above it. That spread is the first thing to keep in mind: "hydrogen water" in a trial is not one product at one dose.
Nothing about a bottle's price, battery capacity, or cycle time in minutes substitutes for the concentration figure. Only the ppm number can be placed against a trial at all.
Unit and general range quoted from the Molecular Hydrogen Institute, "Concentration and Solubility of H2" (molecularhydrogeninstitute.org/concentration-and-solubility-of-h2/), read Sep 7, 2026. Trial concentrations per Sim et al., Sci Rep 2020 (PMID 32699287, full text) and Todorovic et al., Pharmaceuticals 2023 (PMID 37259294, Table 1), read Sep 7, 2026.
Does hydrogen water work? The 2020 Scientific Reports RCT
The most direct answer to "does hydrogen water work" is a specific, named trial, not a general vibe. Sim and colleagues ran a randomized, double-blind, placebo-controlled study in which healthy adults aged 20 to 59 "consumed either 1.5 L/d of HW (n = 20) or plain water (PW, n = 18) for 4 weeks." The hydrogen water measured 0.753 mg/L.
Two results moved. "Apoptosis of peripheral blood mononuclear cells (PBMCs) was significantly less in the HW group" (P = 0.036), and RNA sequencing of those cells found that "transcriptional networks of inflammatory responses and NF-κB signaling were significantly down-regulated in the HW group."
One result did not. The changes in serum biological antioxidant potential, derivatives of reactive oxygen and 8-Oxo-2'-deoxyguanosine "did not differ between groups"; only in participants aged 30 and over did antioxidant potential rise more in the hydrogen group. Stated at its real size: a 38-person trial, four weeks long, with favorable cell-level and gene-expression findings in healthy people, not a symptom improving or a disease outcome changing.
38 healthy adults, 20 to 59 years, 1.5 L/day at 0.753 mg/L or plain water, for 4 weeks.
Significantly less apoptosis (P = 0.036); inflammatory-response and NF-κB networks down-regulated on RNA sequencing.
BAP, reactive-oxygen derivatives and 8-OHdG changes did not differ between groups over the 4 weeks.
Sim M et al., "Hydrogen-rich water reduces inflammatory responses and prevents apoptosis of peripheral blood cells in healthy adults: a randomized, double-blind, controlled trial," Sci Rep 2020 (doi: 10.1038/s41598-020-68930-2, PMID 32699287, PMC7376192), abstract and full text read Sep 7, 2026.
Hydrogen water studies on lipids: the 2023 meta-analysis and the liver-fat trial inside it
The broadest answer to "hydrogen water studies" is a 2023 systematic review and meta-analysis by Todorovic and colleagues, which pooled seven randomized trials and 256 participants in clinical populations. Its result: "a significant reduction in total cholesterol, low-density lipoprotein, and triglycerides after HRW intake (p = 0.01), with small to moderate effects," with pooled standardized mean differences of −0.23, −0.22 and −0.38 respectively.
The seven trials are not one protocol. The pooled table lists people with type 2 diabetes, hypercholesterolemia, overweight middle-aged women, non-alcoholic fatty liver disease (two trials), metabolic syndrome, and elderly adults, on schedules from 4 to 24 weeks, at concentrations listed from about 0.6 ppm up to 15 ppm.
One of those seven is the liver-fat trial: Korovljev and colleagues gave 12 overweight outpatients with mild-to-moderate NAFLD "either 1 L per day of hydrogen-rich water (HRW) or placebo water for 28 days" in a double-blind, placebo-controlled crossover design. On MRI, "HRW significantly reduced liver fat accumulation in individual liver regions-of-interest at 28-day follow-up, as compared to placebo administration (P < 0.05)," a change from 284.0 to 256.5 mM (percent change 2.9%; 95% CI 0.5 to 5.5). Weight and body composition did not differ between treatments, and the authors call their own result "preliminary."
| Endpoint | Result | Where it comes from |
|---|---|---|
| Total cholesterol | Reduced, SMD −0.23 | Pooled across 7 trials, 256 participants, 2023 meta-analysis |
| LDL cholesterol | Reduced, SMD −0.22 | Pooled across 7 trials, 256 participants, 2023 meta-analysis |
| Triglycerides | Reduced, SMD −0.38 | Pooled across 7 trials, 256 participants, 2023 meta-analysis |
| Liver fat (NAFLD, MRI) | Reduced 2.9%, 28 days | One of the 7 pooled trials: 12 patients, 1 L/day, crossover |
Lipid endpoints, populations, durations and concentrations per Todorovic et al., "The Effects of Hydrogen-Rich Water on Blood Lipid Profiles in Clinical Populations: A Systematic Review and Meta-Analysis," Pharmaceuticals 2023 (PMID 37259294, PMC9967957), abstract and Table 1 read Sep 7, 2026. The "about 0.6 ppm" floor is Table 1's own figure for Song 2015; the same paper's discussion text states that "the H2 concentration in HRW in studies included here ranges from 4.76 ppm to 15 ppm", which is the figure our hydrogen water side effects and hydrogen water vs alkaline water pages print, and the two disagree inside the source. Liver-fat result per Korovljev et al., Clin Res Hepatol Gastroenterol 2019 (PMID 30982748), abstract read Sep 7, 2026; that trial is listed in the meta-analysis as one of its seven included studies.
The 2024 systematic review: "Extra Healthy or a Hoax?"
The field's own summary of itself is a 2024 systematic review in the International Journal of Molecular Sciences titled Hydrogen Water: Extra Healthy or a Hoax?, in which two blinded investigators screened the literature and included 25 articles.
"Although preliminary results in clinical trials and studies are encouraging, further research with larger sample sizes and rigorous methodologies is needed to substantiate these findings."
Dhillon et al., Int J Mol Sci 2024, PMID 38256045 / PMC10816294, abstract, read Sep 7, 2026Its conclusion asks for "well-designed studies in humans, with large sample sizes and long-term trials, to ascertain the benefits." Set the trials above against that request and three limits are visible in their own numbers.
- Sample size: 38 adults in the 2020 RCT, 12 patients in the NAFLD trial, 256 participants across all seven pooled lipid trials. The meta-analysis itself notes that "a small number of studies being included in this SRMA ... can impact the strength of observed results and conclusions."
- Duration: 4 weeks in the 2020 RCT, 28 days in the NAFLD trial, 4 to 24 weeks in the pooled trials. The review asks for long-term trials; none here is one.
- Dose: 0.753 mg/L in the 2020 RCT, 4.76 ppm in the NAFLD trial, up to 15 ppm in the pooled trials. Those are different products by concentration, which is why this page keeps returning to the ppm number.
None of that erases the results above. It sizes them correctly: measured, early, and not yet evidence for a broad claim.
Review verdict and conclusions quoted from PMID 38256045 / PMC10816294, read via PubMed and PubMed Central, Sep 7, 2026. Sample sizes, durations and concentrations per the three papers cited above.
Hydrogen water benefits, graded by evidence
Six benefits people ask about, matched to the strongest trial read for this page, or marked as not graded when none was read.
| Benefit | Evidence grade | Trial it rests on |
|---|---|---|
| Inflammatory-response genes and PBMC apoptosis | Early | 2020 Sci Rep RCT, 38 healthy adults, 4 weeks, cell-level |
| Lipids (cholesterol, LDL, triglycerides) | Early | 2023 meta-analysis, 7 trials, 256 participants, small to moderate effects |
| Liver fat | Single small trial | 2019 NAFLD crossover trial, 12 patients, 28 days, 2.9% change (a second NAFLD trial, Kura 2022, ran 8 weeks in the same pooled table) |
| Exercise recovery | Not graded here | No trial read here measured recovery; the 2024 review describes a 2-month trial in juvenile soccer players reporting oxidative-stress and cytokine changes, which we did not read in full |
| Energy and focus | Not graded here | No trial read for this page measured it |
| Skin and anti-aging | Not graded here | No trial read for this page measured it |
Grades reflect only the papers named on this page, read Sep 7, 2026. "Not graded here" means this page read no trial on that endpoint; the 2024 review lists exercise capacity, physical endurance and anti-aging research among areas of ongoing study, so it is not a claim that no such research exists. The second NAFLD trial named in the liver-fat row is Table 1 of Todorovic et al. 2023 (PMID 37259294), which lists "Kura 2022, 30 subjects NAFLD, 3×330 mL HRW (>4 ppm), 8 weeks"; we did not read that trial in full, so it is not graded on its own row.
This is ingredient-level research, not a finished-product test
Every trial here tested hydrogen-dissolved water at a printed concentration; none tested the bottle our own review covers, or any bottle at an unstated ppm.
- The 2020 RCT and the 2023 meta-analysis measured cells, gene expression and blood lipids in study conditions, not how any one person feels day to day.
- The 28-day liver-fat result comes from 12 patients in one crossover trial (Korovljev 2019). It is not the only NAFLD trial in the pool this page read: the same Table 1 lists a second, Kura 2022, in 30 NAFLD patients over 8 weeks, whose liver outcome we did not read in full. That 8-week trial is the liver study our AquaFusion review refers to.
- A bottle that never publishes its own ppm cannot be mapped onto any row above, favorably or unfavorably: there is simply no number to place.
What this means for you: the grades above describe the ingredient across published trials, not any specific product's checkout page.
Why dose is the criterion that decides everything
Every trial cited on this page prints its concentration: 0.753 mg/L in the 2020 RCT, 4.76 ppm in the NAFLD trial, and a table running up to 15 ppm across the pooled lipid trials. Each outcome is reported as the result of drinking water at that stated figure, for that stated number of weeks.
That is what makes concentration the load-bearing fact, not a footnote. A benefit graded "Early" above is early evidence for hydrogen water at the concentration the trial printed, held for the duration it ran.
A bottle without a published ppm figure has not failed the trials; it has removed itself from the comparison. There is no arithmetic that maps an unstated concentration onto a studied one, so skipping either number in this page's two-number check turns a benefit claim into a claim about a different, hypothetical product.
Concentrations per Sim et al. 2020 (full text), Korovljev et al. 2019 (Table 1 of Todorovic et al. 2023) and Todorovic et al. 2023 (Table 1), read Sep 7, 2026.
Who would plausibly notice anything
Reading the trials honestly changes who should expect to notice a difference, and how soon.
Where a signal has been measured
- Healthy adults aged 20 to 59, on PBMC apoptosis and inflammatory-response gene networks, after 4 weeks of 1.5 L/day at 0.753 mg/L.
- Clinical populations in the seven pooled trials (type 2 diabetes, hypercholesterolemia, overweight women, NAFLD, metabolic syndrome, elderly adults), on total cholesterol, LDL and triglycerides.
- Overweight patients with mild-to-moderate NAFLD, on liver fat by MRI, after 28 days of 1 L/day, in one 12-patient trial.
Where this page read nothing
- Anyone expecting an exercise-recovery, energy, focus, skin or anti-aging effect: no trial read for this page measured any of these, though the 2024 review describes a 2-month trial in juvenile athletes we did not read.
- Anyone drinking hydrogen water for under 4 weeks and expecting the liver-fat or lipid results: no positive result on this page came from a trial shorter than 4 weeks.
- Anyone using a bottle with no published ppm: there is no way to know whether it reaches any concentration the trials used.
None of this guarantees any one person will notice anything. It is a map of where researchers measured something, at what dose and window, versus where this page found nothing to read.
Questions people also ask
Does hydrogen water actually do anything, or is it a placebo?
More than a placebo at the biomarker level, less than the marketing implies. A randomized, double-blind, placebo-controlled trial in Scientific Reports (2020) gave 20 healthy adults 1.5 L/day of hydrogen water at 0.753 mg/L for 4 weeks against 18 on plain water: apoptosis of peripheral blood mononuclear cells was significantly less in the hydrogen group, and RNA sequencing found transcriptional networks of inflammatory responses down-regulated, while serum oxidative-stress markers did not differ between groups. A 2023 meta-analysis of seven trials (256 participants) found a significant reduction in total cholesterol, LDL and triglycerides with small to moderate effects. A 2024 systematic review, Hydrogen Water: Extra Healthy or a Hoax?, calls the preliminary results encouraging but says larger sample sizes and rigorous methodologies are needed.
How long do you have to drink hydrogen water before studies saw a change?
No positive result on this page came from a trial shorter than 4 weeks. The 2020 Scientific Reports trial ran 4 weeks; the NAFLD liver-fat trial ran 28 days; and the seven trials pooled in the 2023 lipid meta-analysis ran between 4 and 24 weeks each. A shorter home experiment tests something none of these trials measured.
What concentration of hydrogen did the positive trials use?
The 2020 Scientific Reports trial measured its water at 0.753 mg/L. The 28-day NAFLD trial used 4.76 ppm, and the seven lipid trials pooled in 2023 list concentrations from about 0.6 ppm up to 15 ppm, so the trials do not share one dose. As a general reference, the Molecular Hydrogen Institute states that studies using hydrogen gas dissolved in water range from 0.5 mg/L to 1.6+ mg/L, with most near 1.6 mg/L, and that 1 ppm is about the same as 1 mg/L. A bottle that never states its own ppm output cannot be checked against any of those figures.
Is hydrogen water benefit research funded by bottle makers?
Partly, and the page says so rather than guessing. The 2020 Scientific Reports trial states it was supported by Coway Co., Ltd., and the National Research Foundation of Korea, with the authors declaring no competing interests. The 2019 NAFLD trial lists a co-author affiliated with the Molecular Hydrogen Institute. The 2023 lipid meta-analysis and the 2024 systematic review both declare no conflict of interest. None of the four papers tested the bottle our own review covers.
How we built this page
This is a literature page, not a lab test: we ran no trial and tested no bottle for its own concentration. Every number on the page was checked against the abstract or full text of the paper it is attributed to, on Sep 7, 2026.
- The PBMC apoptosis and inflammatory-response findings, the 0.753 mg/L concentration, the 1.5 L/day and 4-week schedule, the group sizes (20 and 18) and the funding line are from Sim et al., Sci Rep 2020 (PMID 32699287, PMC7376192), abstract and full text.
- The total cholesterol, LDL and triglyceride results (p = 0.01, SMD −0.23, −0.22, −0.38, "small to moderate effects"), the count of 7 trials and 256 participants, and the populations, durations and concentrations of the pooled trials are from Todorovic et al., Pharmaceuticals 2023 (PMID 37259294, PMC9967957), abstract and Table 1.
- The liver-fat result (12 patients, 1 L/day, 28 days, crossover, 2.9% change, 95% CI 0.5 to 5.5) is from Korovljev et al., Clin Res Hepatol Gastroenterol 2019 (PMID 30982748), abstract; the meta-analysis lists it as one of its seven trials.
- The field-level verdict and the calls for "larger sample sizes and rigorous methodologies" and "long-term trials" are from Dhillon et al., Int J Mol Sci 2024 (PMID 38256045, PMC10816294), abstract and conclusions. The general 0.5 to 1.6+ mg/L range and the ppm-to-mg/L equivalence are quoted from the Molecular Hydrogen Institute, molecularhydrogeninstitute.org/concentration-and-solubility-of-h2/.
- We did not evaluate any single commercial bottle's health effects here; a specific checkout is covered separately, where we say so. Nothing here is medical advice, and anyone pregnant, nursing, on medication or managing a diagnosed condition should talk to a doctor first.
Where these numbers actually matter: at checkout
All of the grading above only matters once it meets a real bottle and a real price. That is where the ppm question stops being theoretical.
The objective is simple: run a short, low-cost experiment at a bottle's stated concentration, for at least the 4 weeks the shortest positive trial on this page ran. A 30-day return window at least covers "do I like this," and it covers the 28-day and 4-week windows of the two shortest trials on this page. It does not cover the 8-week NAFLD trial in the same pool, which is why our AquaFusion review calls 30 days a decision window rather than a biomarker test.
Look for a stated ppm or mg/L figure on the product page or box before paying. Find none, and there is nothing to compare against any result on this page.
Set a reminder for 4 weeks out, not 4 days. Every positive result graded above needed at least 4 weeks, not a single glass, before anything was measured.
Affiliate disclosure: the review linked below contains affiliate links, and a purchase made through them may earn us a commission at no extra cost to you. This page carries no affiliate link of its own.
Read our AquaFusion reviewThe cheapest way to run your own 30-day experiment: real checkout prices, and why the one number it does not publish is the ppm this page just taught you to look for.