GLP-1 is the hormone behind the most talked-about class of weight-loss drugs in a generation, so a trial reporting that hydrogen-rich water raised plasma GLP-1 levels is going to attract attention[1]. It is worth being precise about what that trial did and did not show, because the gap between “increased a hormone” and “caused weight loss” is where most supplement marketing lives.
What GLP-1 does
Glucagon-like peptide-1 is a hormone released from the gut in response to food. It slows how quickly the stomach empties, promotes insulin release when blood glucose is elevated, and acts on brain circuits involved in satiety. Those are the mechanisms that GLP-1 receptor agonist drugs exploit, at pharmacological doses far above what the body produces on its own.
That last point is the one to hold onto while reading the rest of this page. A modest shift in circulating GLP-1 is not the same intervention as a drug engineered to saturate the receptor for a week at a time. The mechanism being shared does not make the effects comparable.
The HYDRAPPET trial
The trial was a randomized, placebo-controlled, double-blind study named HYDRAPPET[1]. It included 36 participants, 24 of them female, with a mean age of 42.1 years and a mean body mass index of 30.8, which places the group in the obese range.
Participants were randomized to either 1.0 litre of hydrogen-rich water containing 15 milligrams of hydrogen daily, or 1.0 litre of control water containing zero hydrogen, for eight weeks[1]. Matching the fluid volume between arms is the correct design choice and rules out the hydration confounder that weakens some other hydrogen trials.
The trial is registered at ClinicalTrials.gov under NCT06722326. Pre-registration matters because it makes it harder to quietly reframe which outcome was the primary one after seeing the data.
The results
Compared with control water, hydrogen-rich water significantly mitigated cravings at p equal to 0.05, improved subjective sleep quality at p equal to 0.05, reduced total cholesterol at p equal to 0.02, reduced LDL cholesterol at p equal to 0.04, and increased plasma glucagon-like peptide-1 levels at p equal to 0.05[1]. No severe adverse effects were reported.
Read that list again and note what is missing from it. The trial measured body composition. Body composition is not among the outcomes reported as significantly different. Cravings changed, a satiety hormone changed, and lipids changed, over eight weeks, in 36 people. Body weight and body composition did not separate between the groups.
That is the single most important thing on this page. A trial that measured appetite, hormones and body composition, and found effects on the first two but not the third, is evidence about appetite regulation. It is not evidence about weight loss, and presenting it as such would be a misreading of the authors’ own results.
Three caveats the trial itself flags
First, the p-values. Three of the five reported outcomes sit exactly at p equal to 0.05, which is the boundary of the conventional significance threshold rather than a comfortable distance inside it. A result at exactly 0.05 is the weakest form of a positive finding. It means that if there were no true effect, you would still see a difference this large roughly one time in twenty.
Second, the sample size. Thirty-six participants split across two arms is 18 per group. Small trials produce noisier estimates, and when several borderline p-values appear together in a small trial, the probability that at least one is a false positive rises accordingly.
Third, and stated plainly in the paper’s own conflict-of-interest declaration, one author is employed by the company that supplied the product tested[1]. Industry involvement does not invalidate a result, and disclosing it is what integrity looks like. It is a reason to weight independent replication heavily.
How it fits with the longer trials
Two other trials give useful context on the metabolic side.
A 24-week randomized, double-blinded, placebo-controlled trial in 60 people with metabolic syndrome tested high-concentration hydrogen-rich water delivering more than 5.5 millimoles of hydrogen per day[2]. It found significant reductions in blood cholesterol and glucose, attenuated haemoglobin A1c, and improved biomarkers of inflammation and redox homeostasis against placebo, at p less than 0.05.
On body composition specifically, that trial reported that hydrogen “tended to promote” a mild reduction in body mass index and waist-to-hip ratio[2]. The word “tended” is doing real work there. It is the standard way of describing a change that moved in the expected direction without reaching statistical significance. Twenty-four weeks, 60 people, high-concentration dosing, and body composition still did not clearly separate.
A six-month trial in adults aged 70 and over, using 15 parts per million hydrogen-rich water, similarly reported no significant differences between interventions for most outcomes including body composition, with the exception of an improved chair-stand performance at p equal to 0.01[3].
Three trials, running 8, 24 and 26 weeks, all measuring body composition, and none of them showing a clear body-composition effect. The cholesterol finding, by contrast, appears in two of the three independently, which makes it the more replicated result.
What the cravings finding might mean
Cravings were measured as a patient-reported outcome, and self-reported appetite in an unblinded-to-purpose participant is a soft endpoint even inside a double-blind trial. The trial was double-blind, which is the right protection, and hydrogen water is genuinely difficult to distinguish from plain water, so the blinding is more credible here than in a trial comparing a pill to nothing.
If the cravings effect is real, the GLP-1 finding offers a coherent story for why: more circulating satiety hormone, less craving. Coherence is not proof. The trial did not test whether the GLP-1 change caused the craving change, and with 18 people per arm it was not powered to.
The practical implication of a cravings effect without a body-composition effect is narrow but not nothing. Feeling less driven to snack has value even if the scale does not move, and eight weeks is short enough that a genuine appetite change might not have had time to show up as measurable body composition change.
What this does not support
Hydrogen water is not a GLP-1 drug and does not belong in the same conversation as one. The drugs produce sustained pharmacological receptor activation and have large randomized trials with weight loss as the primary endpoint. This trial produced a borderline change in an endogenous hormone with no accompanying body-composition change in 36 people over eight weeks.
Anyone considering hydrogen water as a substitute for prescribed metabolic treatment should not. Anyone considering it alongside one should mention it to the clinician managing that treatment.
The bottom line
The HYDRAPPET trial found that eight weeks of hydrogen-rich water at 15 milligrams of hydrogen daily reduced cravings, improved subjective sleep quality, lowered total and LDL cholesterol, and raised plasma GLP-1, in 36 obese adults[1]. Several of those results sit exactly at the p equals 0.05 threshold, the sample is small, and one author works for the product supplier.
The cholesterol effect is the finding with independent support from a longer, larger trial[2]. The body-composition effect is the one that did not appear in this trial, or in the 24-week trial, or in the six-month trial. Treat the GLP-1 result as an interesting mechanistic lead and not as a weight-loss claim.
References
- The Effects of 8-Week Hydrogen-Rich Water Consumption on Appetite, Body Composition, Sleep Quality, and Circulating Glucagon-like Peptide-1 in Obese Men and Women (HYDRAPPET): A Randomized Controlled Trial. Medicina (Kaunas), 2025
- The Effects of 24-Week, High-Concentration Hydrogen-Rich Water on Body Composition, Blood Lipid Profiles and Inflammation Biomarkers in Men and Women with Metabolic Syndrome. Diabetes, Metabolic Syndrome and Obesity, 2020
- The effects of 6-month hydrogen-rich water intake on molecular and phenotypic biomarkers of aging in older adults aged 70 years and over: A randomized controlled trial. Experimental Gerontology, 2021
These statements have not been evaluated by the Food and Drug Administration. This information is not intended to diagnose, treat, cure, or prevent any disease. Content is for informational purposes only and is not medical advice; consult a qualified healthcare provider before starting any supplement. As an Amazon Associate we earn from qualifying purchases.
