Effect of hydrogen/oxygen therapy for ordinary COVID-19 patients: a propensity-score matched case-control study
Zeng Y, Guan W, Wang K, Jie Z, Zou X, Tan X, Li X, Chen X, Ren X, Jiang J, Zheng Z, Shi J, Zhong N. · BMC Infectious Diseases. 2023;23:440.
What kind of evidence is this?
Human
Three-centre retrospective propensity-score-matched case-control study
Respiratory health
Inhaled H₂
Not reported in this record.
Not reported in this record.
Methods at a glance
Adults hospitalized in China with ordinary, non-severe COVID-19 between December 2019 and March 2020.
180 eligible records before matching: 42 H₂/O₂ and 138 oxygen. The principal 1:2 matched analysis included 33 H₂/O₂ and 55 oxygen patients.
At least six hours/day for more than seven consecutive days; clinical follow-up through discharge and day-28 discharge-rate analysis.
Nasal-mask H₂/O₂ for at least six hours daily on more than seven consecutive days, plus standard care.
Brown's gas / oxyhydrogen at an H₂:O₂ volume ratio of 2:1.
Total mixture flow was 3.0 L/min; direct 2:1 calculation gives H₂ flow 2.0 L/min and O₂ flow 1.0 L/min.
2,000 mL/min.
1,000 mL/min.
Standard care plus low-flow air/oxygen therapy below 3.0 L/min, with duration adjusted clinically.
Outcomes and reported result
Hospital length of stay, discharge by days 14, 21 and 28, day-5 SpO₂, vital signs, respiratory symptoms, adverse events and mortality.
Matched median stay was 12 versus 13 days and later discharge rates and day-5 SpO₂ favored H₂/O₂. Day-14 discharge did not differ significantly. Gender and time from symptom onset were not associated with stay within the H₂/O₂ subgroup; no significant adverse effects were observed.
The complete free PMC article and supplementary-analysis descriptions were checked for the unmatched and matched cohorts, gas ratio/flow, schedule, primary and null results, safety, funding and conflicts.
A reported association, difference or mechanism is not automatically a clinical benefit. Null findings, outcome type, study design and precision all matter.
Cautions and applicability
Retrospective nonrandomized study with small matched groups, propensity matching on only age, sex and time to admission, possible unmeasured confounding, incomplete post-treatment laboratory/SpO₂ data and early-pandemic care. Public Chinese grants were reported and authors declared no competing interests.
Ordinary COVID-19 hospitalized during the original 2019–2020 outbreak; contemporary treatment relevance is uncertain.
Not reported in this record.
Not reported in this record.
Not reported in this record.
No single-study GRADE certainty rating is assigned. Certainty is assessed by important outcome across a complete eligible evidence set, not by attaching a final grade to one source.
Sources and record status
PMID: 37386364 · DOI: 10.1186/s12879-023-08424-4
Free full article in PubMed Central.
Complete PMC article, supplementary-analysis descriptions and PubMed metadata; full-text extraction checked 9 August 2026.
2026-08-10