Hydrogen jelly did not improve sleep overall

Molecular hydrogen is marketed on the back of reported antioxidant and anti-inflammatory effects, and a growing number of products promise better rest and lower stress. The authors point out that evidence for any effect on sleep quality in healthy people remains extremely limited. A second, less obvious question sits underneath: gut bacteria themselves produce hydrogen, so two people taking the same supplement may not be starting from the same internal hydrogen exposure. Whether that inter-individual variability blunts or amplifies a supplement's effect has seldom been studied, and it matters for how any such trial should be read.

This was a randomised, double-blind, placebo-controlled, parallel-group trial. A total of 44 healthy adults who reported poor sleep quality were allocated to receive either a hydrogen-rich jelly or a matching placebo jelly. Sleep and mood outcomes were captured with several validated instruments: the OSA-MA sleep questionnaire, a visual analogue scale (VAS) for sleep quality, the Pittsburgh Sleep Quality Index (PSQI) and the State-Trait Anxiety Inventory (STAI). Gut microbiota composition was also profiled, allowing the researchers to look at β-diversity and at the relative abundance of hydrogen-producing bacteria such as Bacteroides, then test group-by-microbiome interactions using linear mixed-effects models.

The headline result was null: across the whole sample there were no significant differences between the hydrogen and placebo groups in changes to OSA-MA, VAS, PSQI or STAI scores. The interesting signal came from the secondary, exploratory analyses. When participants were split by their mean change in the VAS score for sleep quality, gut microbiota β-diversity showed apparent clustering in the hydrogen group but not in the placebo group, and that clustering tracked differences in the abundance of hydrogen-producing bacteria such as Bacteroides. Mixed-effects models found significant group-by-hydrogen-producer interactions for both sleep-quality VAS and mental stress, with greater improvement among those carrying fewer hydrogen-producing bacteria.

The overall null adds a well-controlled negative result to a thin and largely promotional evidence base around hydrogen and sleep, which is useful in itself. The microbiome finding is a hypothesis rather than a conclusion — the authors frame it exactly that way, as an idea warranting confirmation in larger studies and as a case for microbiome-informed stratification in future hydrogen research. Open questions are substantial: the interaction was found after stratifying on an outcome, the trial was small, only one product format and population were tested, and it remains unknown whether measured hydrogen-producing capacity would predict response if specified in advance.

If you are considering a hydrogen product for sleep, the honest reading is that in this trial it did nothing measurable to sleep or anxiety on average. That is a reason to be sceptical of confident marketing claims about hydrogen and rest. The subgroup result is a reminder that supplement responses may genuinely differ between people for biological reasons, but a signal found in exploratory analysis of a small trial is not a basis for predicting your own response. Nobody can currently tell you whether your gut bacteria make you a responder.