This small randomized trial did not detect an advantage for its oral reduced-glutathione capsules over placebo on the reported four-week oxidative-stress and red-blood-cell glutathione comparisons. It studied healthy adults and laboratory biomarkers, not symptom relief, disease prevention or treatment benefit.[1]
The result needs two qualifications: one randomized participant was omitted from the final analysis, and several descriptions of the measurements and results conflict within the indexed article. This record preserves those uncertainties. It is based on accessible indexed primary text, not a fresh complete paper, original analysis plan or visual audit of the figures.
The Study Library keeps this experiment separate from compound overviews and other glutathione studies.
Study at a glance
| Field | What the reviewed text reports |
|---|---|
| Publication | Allen and Bradley, 2011; PMID 21875351; PMCID PMC3162377 |
| Design | Randomized, double-blind, placebo-controlled trial at Bastyr research and clinical sites |
| Population | 40 healthy adults; baseline table labels 20 in each arm |
| Test article | Oral reduced L-glutathione capsules |
| Comparator | Matched placebo capsules |
| Follow-up | Four weeks; 39 participants completed |
| Published analysis hierarchy | Three primary biomarker endpoints in the Data Analysis section; the abstract emphasizes two urinary primary outcomes |
| Main finding | No statistically significant active-versus-placebo difference on the featured biomarker comparisons |
| Transparency limit | No registration identifier was seen in the indexed text reviewed; no registry search or original protocol review was performed |
These fields identify the study, not a practical-use regimen. The healthy-volunteer setting also limits what it can say about people with a diagnosed illness, different baseline nutritional status or concurrent treatments.
What the study measured, and what it did not
The urinary measures were F2-isoprostanes and 8-hydroxy-2′-deoxyguanosine, abbreviated 8-OHdG. The red-blood-cell measures concerned reduced glutathione, GSH, oxidized glutathione, GSSG, and their ratio. They address aspects of biochemical status and oxidative stress.[1]
The published Data Analysis section calls the two urinary measures and the erythrocyte GSH:GSSG ratio its three primary endpoints. It describes other erythrocyte measures and adverse events as secondary. The abstract instead names the two urinary primary outcomes. This page follows the explicitly stated analysis hierarchy while keeping the abstract's narrower wording visible.
That is a description of the publication, not independent confirmation that all three endpoints and their analysis rules were fixed before enrollment. A dated protocol, statistical analysis plan and registration history were not obtained.
None of these measurements directly establishes improved daily functioning, symptoms, quality of life or disease outcomes. Even a favorable laboratory change would require further evidence before being treated as patient benefit. The biomarker and clinical-outcome guide explains why the measured outcome matters.
The reported biomarker comparisons
The table keeps primary and additional measures separate. The ± quantities reproduce the publication’s notation; they are not confidence intervals for the between-group treatment difference.[1]
| Outcome | Active versus placebo as reported | Comparison |
|---|---|---|
| Urinary F2-isoprostanes; primary in Data Analysis | Mean change 0.0±0.1 versus 0.0±0.1 ng/mg creatinine | P=.38 |
| Urinary 8-OHdG; primary in Data Analysis | Mean change −0.2±3.3 versus +1.0±3.2; disputed denominator unit omitted | P=.27 |
| Erythrocyte GSH:GSSG ratio; primary in Data Analysis | Raw values omitted because the result label is ambiguous | P=.99 |
| Reduced erythrocyte glutathione; additional measure | Mean change −74.5±294.8 versus −34.5±334.7 µmol/L | P=.67 |
| Total erythrocyte glutathione; additional measure | No significant between-group change reported in Figure 3 caption | P=.70 |
| Oxidized erythrocyte glutathione; additional measure | No significant between-group change reported | Reported only as P>.05 |
The two F2-isoprostane means are the same at the displayed precision; that is compatible with a nonidentical underlying comparison. It is not a reason to replace the reported P value or conclude that every participant had the same response.
None of these tests demonstrated the proposed biomarker advantage. But nonsignificance is not an equivalence test. This review did not establish dependable between-arm confidence intervals or an equivalence margin. It therefore cannot say that all smaller effects were excluded. See confidence intervals, P values and multiplicity for the distinction between an estimate, its uncertainty and a decision threshold.
Reporting inconsistencies that remain unresolved
Several details prevent a more confident reconstruction of the results.[1]
First, the abstract and Results use different creatinine denominator units for 8-OHdG: micrograms per gram versus micrograms per milligram. The numerical changes and P=.27 can be reported as printed, but choosing one unit would introduce an unverified correction.
Second, the Results calls the negative GSH:GSSG values “final ratios,” while the baseline table reports positive ratios. The available text does not justify silently relabeling those negative numbers as changes. The table above therefore preserves P=.99 without treating the raw numbers as verified final concentrations, ratios or changes.
Third, the abstract describes tandem mass spectrometry for glutathione, whereas the Methods describes an enzymatic recycling assay. A Figure 1 caption is headed 8-OHdG but then refers to urinary GSH. That caption inconsistency is not evidence of a separate urinary-glutathione endpoint.
The baseline table's female-percentage row also conflicts with the participant totals described in the text. No corrected sex distribution by arm is reconstructed here. These are reporting limits, not proof that the experiment was invalid. Original figures, laboratory documentation or an authoritative correction could clarify them; they were not available in this review.
One randomized participant was omitted
The article reports 39 completers and one woman lost after baseline. The authors say they omitted her baseline data and did not estimate her final outcomes because they were addressing mechanistic efficacy.[1]
Her randomized arm is not identified in the reviewed account. The baseline table's 20-per-arm headings do not establish that all 40 participants supplied four-week outcomes. Nor do they justify assigning 19 observed outcomes to a particular arm.
The final account therefore excludes a randomized participant; it should not be presented as an analysis of all randomized participants with observed or imputed outcomes. The effect of that exclusion cannot be determined from the available text. Neither the small number missing nor the direction of the reported results proves that the omission was inconsequential.
The intention-to-treat and analysis-set guide explains why randomized, followed-up and analyzed populations are not interchangeable.
Analysis, diet and the limits of a null result
The published analysis describes checking distributions and using analysis of variance and paired t-tests to examine changes. This review did not verify a complete multiplicity-control plan, allocation-sequence and concealment procedures, or an original prospectively frozen statistical plan.[1]
Baseline dietary estimates differed for cysteine, glutamic acid and protein, each with P=.04; methionine was P=.06. The authors report no association between baseline glutathione and dietary factors, with supporting data not shown. That statement does not establish that diet imbalance could not influence the comparison. Baseline P values also cannot, by themselves, certify or disprove proper randomization.
The discussion describes 80% power for a twofold reduction in 8-OHdG. Power to detect a large change is not proof that false-negative error was impossible, nor that smaller effects were precisely excluded. This page does not adopt the discussion's stronger argument about ruling out that error.
The appropriate conclusion is limited: this small, short study did not demonstrate the tested biomarker effects. It did not establish universal ineffectiveness, prove that intact oral glutathione cannot be absorbed, or determine what would happen at other assessment times. The null-results guide gives the broader interpretation.
Adverse events and adherence
The article describes adverse-event questionnaires at baseline and weeks two and four, alongside laboratory monitoring. It reports no serious or systematic adverse effects and no routine hematology, liver or kidney laboratory changes during the study.[1]
| Mild reports described across the study | Reported number |
|---|---|
| Increased flatulence or loose stools | 5 |
| Flushing | 2 |
| Weight gain | 1 |
The report does not provide a dependable arm assignment or establish whether those categories overlap. They must not be summed into eight distinct affected people, assigned to active treatment, or converted into comparative safety rates.
The absence of a reported serious event in a small healthy-volunteer trial cannot exclude uncommon or delayed harm. It also cannot establish safety in other populations or with other formulations and routes.
Interviews and pill counts were used to assess adherence; the authors report more than 90% adherence in both groups. This is a reported adherence assessment, not independent verification of ingestion, systemic exposure or perfect compliance. It does not resolve the missing outcome or the other analysis limitations.
Funding and the identity of the tested material
Kohjin Co Ltd of Japan funded the trial and supplied the reduced glutathione. Natural Factors encapsulated the active material and placebo. The authors also reported partial support through NIH programs and declared no competing financial interests.[1]
The no-competing-interest declaration does not erase the separately disclosed commercial funding and material support. Conversely, sponsorship alone does not determine whether a result is credible. The methods, reporting completeness and access to evidence remain important.
Historical manufacturing and purity statements in a paper are not an independently authenticated certificate for a current product. This experiment cannot validate a catalog material or establish equivalence to a different capsule, liposomal preparation, precursor or nonoral formulation. The drug-versus-research-material guide explains that identity boundary.
A separate six-month oral-glutathione trial has its own population, follow-up and analysis. Its findings should not be substituted for this experiment or pooled informally to repair this paper's unresolved reporting details.
Frequently asked questions
Did the trial demonstrate a biomarker benefit?
No statistically significant advantage over placebo was reported for the three endpoints identified as primary in the published Data Analysis section. That is not proof of equivalence or of no possible effect.
Were all 40 randomized participants analyzed?
No. The report describes 39 completers and omission of one participant lost after baseline. Her arm and a dependable endpoint-by-arm follow-up breakdown were not established.
Why are some result units or values omitted?
The text gives conflicting units for 8-OHdG and an ambiguous label for negative ratio values. Omitting those disputed details is more accurate than silently choosing a correction.
Does the result prove that oral glutathione is not absorbed?
No. The study's fasting biomarker comparisons were not a direct test of intact absorption kinetics, and a null comparison cannot establish that broader mechanism.
Did the study demonstrate clinical benefit?
No. It measured biomarkers in healthy volunteers, not symptoms, disease prevention, quality of life or treatment outcomes.
Were the capsules proven safe?
No. Short-term adverse-event and laboratory observations were reported, but mild-event arm assignments were unclear and the study could not exclude uncommon or longer-term harm.
Was registration or correction status fully checked?
No. No registration identifier was seen in the indexed text reviewed. Direct article and PubMed access encountered CAPTCHA, and no registry search or comprehensive notice review was performed.
Primary-source links
- Primary article: PMC3162377; indexed text reviewed, not a fresh complete article or figure-image audit
- PubMed identifier: PMID 21875351; direct record access was challenged
- DOI identifier: 10.1089/acm.2010.0716; not newly resolved in this review
Review scope and corrections
Editorial evidence review: September 3, 2026, in America/Edmonton and UTC. Serum Research is responsible for this source-based educational summary. The review covered indexed primary abstract, methods and analysis text, results, Table 1 text, figure captions, discussion and disclosures. It did not include a fresh complete paper/PDF, actual figure or table images, original protocol or statistical plan, participant data, registration history or a comprehensive correction/retraction check. No independent clinical or statistical specialist review is claimed.
Return to the Research Library. To report a source discrepancy, use Contact Us.