2020 Journal of thoracic disease Pilot / Observational Human H₂ therapy Inhalation Respiratory / COVID
2020 · Guan et al. — Hydrogen/Oxygen Mixed Gas Inhalation Improves Disease Severity and Dyspnea in Patients with Coronavirus Disease 2019 in a Recent Multicenter, Open-Label Clinical Trial
Super-Abstract
Among patients with COVID-19 and breathlessness, more patients who inhaled a hydrogen-oxygen mixture on top of standard care improved in disease severity by the end of treatment than controls (70.5% compared with 31.8%, 44 analysed per group), but clinicians, not chance, decided who received the gas. The share of patients whose breathlessness improved was higher with the gas at day 2; at day 3 and at the end of treatment this comparison was not significant, while the authors report significant differences when symptom scores were analysed as continuous values. Without randomisation or blinding, selection bias may explain part of the difference, as the authors acknowledge. (Journal of Thoracic Disease, 2020.)
Commentary
The authors proposed that a hydrogen-oxygen mixture might ease breathing in COVID-19 because it meets less resistance in the airways than room air. Between January 21st and March 23rd, 2020, seven hospitals in China enrolled patients aged 18 to 85 with laboratory-confirmed COVID-19 and dyspnea both on admission and at enrollment. Randomisation was not used: attending clinicians assigned patients to the treatment or control group. In addition to standard care, the treatment group inhaled 66% hydrogen and 33% oxygen through a nasal cannula daily until discharge, at 3 L/min (corrected from 6 L/min in a 2020 erratum), for a median of 7.7 hours a day. Controls received standard care with daily oxygen therapy, for a median of 24 hours a day. Of the 633 screened, 44 were included in the treatment group and 46 in the control group. The primary endpoint was the share of patients whose disease severity improved by at least one category on a five-category scale.
Key quotes
- „The therapeutic effects of H2-O2 became significant as early as days 2 and 3 and the amelioration of most respiratory symptoms persisted till the end-of-treatment“ — the authors' summary; their table shows no significant end-of-treatment difference for dyspnea or cough
- „We neither randomly assigned patients with COVID-19 due to the emergency nor matched the patients with propensity scores, which could have resulted in selection bias.“ — the authors' own statement of the main weakness of the design
- „Nonetheless, H2-O2 inhalation might be considered useful to patients with dyspnea or those in facilities without sufficient oxygen supplies.“ — the authors' closing view, written after listing these limitations
Our assessment
The trial reports large differences, but its design limits what they can show. Because clinicians chose who received the gas and the trial was open-label, the groups may have differed in ways that shaped the results. The authors themselves name selection bias, the open-label design and the variable duration of inhalation as limitations. Exposure also differed: the hydrogen-oxygen mixture was given for a median of 7.7 hours a day, oxygen in the control group for 24 hours. Limitations: no randomisation, blinding or propensity matching, outcomes rated on symptom scales, an empirically set protocol, and many endpoints tested at several time points. The text states that chest distress and chest pain improved with all P values below 0.05, but its table shows no significant difference for chest distress at the end of treatment. The discussion calls the study randomised, which contradicts the methods. No serious adverse events were reported.
Study design
- Type: open-label, multicentre, non-randomised controlled trial (seven hospitals in China, January to March 2020), allocation by attending clinicians · n: 44 H₂-O₂ and 46 control enrolled (primary endpoint analysed in 44 and 44), of 633 screened · H₂ delivery: inhaled 66% hydrogen and 33% oxygen at 3 L/min (corrected by erratum) via nasal cannula, daily until discharge, median 7.7 hours a day and 64 hours in total
- Comparator: standard care with daily oxygen therapy (median 24 hours a day), allocation by clinicians
- Result: primary endpoint (improved disease severity): day 2 20.5% vs. 2.3% (P=0.019), day 3 31.8% vs. 11.5% (P=0.038), end of treatment 70.5% vs. 31.8% (P<0.001). Dyspnea improved more at day 2 (50.0% vs. 23.9%, P=0.019) the categorical comparison was not significant at day 3 or at the end of treatment, while continuous analyses in the supplement were reported as significant. Cough improved more at days 2 and 3, chest pain and resting oxygen saturation at all time points. Adverse events: worsening cough 6.8% vs. 8.7%, chest distress 2.3% vs. 21.7%, no serious adverse events
Source & links
Screenshot of the PubMed page
This page mirrors the published abstract (© the authors / publisher) for reference and citation. The canonical source is the PubMed record linked above. This is not medical advice.