Hydrogen-oxygen therapy alleviates clinical symptoms in twelve patients hospitalized with COVID-19: A retrospective study of medical records.

ABSTRACT: A global public health crisis caused by the 2019 novel coronavirus disease (COVID-19) leads to considerable morbidity and mortality, which bring great challenge to respiratory medicine. Hydrogen-oxygen therapy contributes to treat severe respiratory diseases and improve lung functions, yet there is no information to support the clinical use of this therapy in the COVID-19 pneumonia.A retrospective study of medical records was carried out in Shishou Hospital of Traditional Chinese Medicine in Hubei, China. COVID-19 patients (aged ≥ 30 years) admitted to the hospital from January 29 to March 20, 2020 were subjected to control group (n = 12) who received routine therapy and case group (n = 12) who received additional hydrogen-oxygen therapy. The clinical characteristics of COVID-19 patients were analyzed. The physiological and biochemical indexes, including immune inflammation indicators, electrolytes, myocardial enzyme profile, and functions of liver and kidney, were examined and investigated before and after hydrogen-oxygen therapy.The results showed significant decreases in the neutrophil percentage and the concentration and abnormal proportion of C-reactive protein in COVID-19 patients received additional hydrogen-oxygen therapy.This novel therapeutic may alleviate clinical symptoms of COVID-19 patients by suppressing inflammation responses.

Introduction

In December 2019, a novel coronavirus named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) caused acute pneumonia disease (COVID-19) broke out in Wuhan, China, and quickly spread to many countries worldwide, especially in Europe, America, and South-East Asia.[ By May 1, 2021 globally, over 153 million cases and more than 3.2 million deaths with COVID-19 disease have been confirmed globally from World Health Organization (WHO) situation report.[ Compared with SARS, which also spread around the world in 2003, SARS-CoV-2 was more contagious. Compared with flu, SARS-CoV-2 led to a higher mortality rate.[ On January 30, 2020, the WHO declared the COVID-19 pneumonia as a “Public Health Emergency of International Concern” , and on March 12, it officially declared COVID-19 a global pandemic, this means not only an escalation in the severity of the epidemic, but also an escalation in the difficulty of fighting it.[ Due to the lack of medicines and vaccines, basic health services for COVID-19 are often interrupted. The number of COVID-19 deaths increases with age especially over 80 years old; moreover, basical diseases such as cardiovascular, cerebrovascular, and respiratory diseases increase the risk of death for COVID-19 patients and lead great difficulties to treatment.[ Therefore, there is an urgent need for effective therapeutics to not only reduce the infection and fatality rates of COVID-19, but also improve the survival status of patients.

Although Remdesivir, Chloroquine, Hydroxychloroquine, etc, have proven to be effective for the treatment of COVID-19, clinical data indicates that these small molecule drug candidates exhibited multiple side effects, for instance, headache, dizziness, diarrhea, stomach cramps, vomiting, arrhythmia, low blood pressure, liver inflammation, and sweating.[ Hypoxia initiated by lung injury leads to multiple organ damage[; thus, alleviating respiratory distress can reduce the risk of death in severe COVID-19 patients. Combined with anti-inflammatory and potential viricidal properties and relative high security, oxygen therapy has been reported to reverse the hypoxia caused by the SARS-CoV-2 virus.[ Recently, molecular hydrogen (H2) has been applied as therapeutic medical gas to reduce various organ system dysfunctions and improve survival rate due to its anti-oxidant, anti-inflammatory, and anti-apoptotic actions. It has been found that inhaled hydrogen therapy ameliorated acute lung injury in animal models[; hence, we proposed a hypothesis that hydrogen-oxygen therapy may be beneficial to COVID-19 patients in protecting organs from hypoxia-induced damage.

Hydrogen-oxygen nebulizer therapy, an interesting new strategy that has been officially recommended by the New Coronavirus Pneumonia Prevention and Control Program (7th edition) published by the National Health Commission of China,[ is of great significance to verify its validity. A multicenter, randomized, double-blind, parallel-group controlled trial has shown that hydrogen-oxygen therapy is superior to oxygen therapy in terms of safety and tolerability in patients with acute exacerbation of chronic obstructive pulmonary disease .[ Oxygen therapy and noninvasive mechanical ventilation have limited effect on the emergency treatment of tracheal stenosis,[ and also have important application in the treatment of critical respiratory failure in intensive care unit. Low density and low molecular weight gaseous helium (He) has been shown to reduce airway resistance and inspiratory force in patients with upper airway obstruction.[ However, due to the high cost of helium storage and production, it cannot be widely used in clinic. So far, in the emergency treatment of tracheal stenosis, there is no other way to reduce the inspiratory capacity satisfactorily except He inhalation. H2 has similar physical properties to He, so it may reduce the airway resistance of narrow trachea. In addition, hydrogen can be produced easily and cheaply by electrolyzing water, it can even be available in ambulance.[ Hydrogen generator is new equipment, which has been reported to be used in the deposition of calcium oxalate in kidney, cerebral ischemia-reperfusion injury and chronic obstructive pulmonary disease model rats, and it showed good effects in reducing inflammation and biological safety.[ However, up to now, the clinical application of hydrogen-oxygen therapy in COVID-19 patients has not been reported.

In this study, we retrospectively collected and analyzed the detailed clinical data of SARS-CoV-2-infected patients diagnosed in the laboratory of Shishou Hospital of Traditional Chinese Medicine (Jingzhou, Hubei). In this study, we analyzed the clinical characteristics of COVID-19 pneumonia patients, evaluated the effectiveness and safety of hydrogen-oxygen mixture generated by medical hydrogen-oxygen nebulizers in the treatment of COVID-19, in viewing to provide a new clinical therapy for the prevention and treatment of this major infectious disease.

Materials and methods

Study design and patients

We did a retrospective review of medical records from 24 adult patients with moderate and severe COVID-19 pneumonia admitted to Shishou Hospital of Traditional Chinese Medicine from January 29 to March 20, 2020. COVID-19 is diagnosed according to the New Coronavirus Pneumonia Prevention and Control Program (7th edition) published by the National Health Commission of China.[ The definition of severe patients is that adult meets any of the following items: shortness of breath, respiratory rate > 30 times min−1; resting state, oxygen saturation is less than 93%; arterial oxygen partial pressure /oxygen inhalation is less than 300 mm Hg (1 mm Hg = 0.133 kPa); pulmonary imaging showed that the lesions progressed more than 50% in 24 to 48 hours.

The patients recruited in this study meet any of the following criteria: patients were diagnosed as COVID-19; patients were aged between 18 and 80 years; patients were treated with nebulized hydrogen-oxygen mixture. Exclusion criteria: patients who have withdrawn from hydrogen-oxygen therapy; patients accompanied with liver, kidney, and heart diseases.

This study was reviewed and approved by the Medical Ethical Committee of Affiliated Hospital of Guangdong Medical University (Ethical approval No. PJ2020-029). Written informed consent was obtained from each enrolled patient.

Data collection

Clinical records, laboratory findings, and chest computed tomography (CT) scans were reviewed for 24 patients hospitalized with COVID-19. All information is obtained and managed through custom data collection tables, and 2 researchers independently collected and reviewed data collection tables to verify the accuracy of the data.

Basic clinical information of new admitted patients was collected, including gender, age, smoking history, other disease histories. Also, whether patients have direct contact with other COVID-19 patients were investigated. The interval time of CT examination, clinical symptoms including fever, cough, expectoration, fatigue, sore throat, diarrhea, and the results of the first and last laboratory examinations including immune inflammation indicators, electrolytes, myocardial enzyme profile, liver function, and renal function as well as the clinical status, cured, discharged, in treatment, critical, and death up to the study date was analyzed.

CT examination method

Chest plain scan was performed with United imaging 40 row CT instrument. The scanning parameters were: thickness 2.00 mm, cols 512, rows 512, zoom 1.00, window level 600, window width 1600. The patient was in supine position. The scanning range was from the apex to the costophrenic angle at the bottom of the lung. Two senior radiographers in the cardiothoracic group reviewed the films, analyzed the distribution, scope, shape, pleural effusion, lymph node enlargement of the patients, and analyzed the imaging manifestations of different clinical types and the relationship with the course of disease. To facilitate the statistics, the left and right lungs were divided into 10 segments, a total of 20 segments.

Laboratory diagnosis

According to WHO guidelines and regulations of the Chinese National Health Commission, throat swab samples of COVID-19 patients were collected, and quantitative RT-PCR recommended by China Center for Disease Control and Prevention was used to detect SARS-CoV-2. All laboratory work was completed by staff from Da An Gene Co, Ltd. of Sun Yat-Sen University and COVID-19 positive case was certified according to the New Coronavirus Pneumonia Prevention and Control Program (7th edition) published by the National Health Commission of China.[

Hydrogen-oxygen therapy

All the recruited patients received antiviral and symptomatic treatment, and 12 of them were additionally inhaled with hydrogen-oxygen mixture produced by the Medical Hydrogen-Oxygen Generator with Nebulizer (AMS-H-03, Asclepius Meditec Co, Shanghai, China), which was listed by the State Food and Drug Administration in national innovation 3 categories of respiratory medical equipment. As shown in previous reports, the Medical Hydrogen-Oxygen Generator with Nebulizer that has been approved for public use can provide an output for the hydrogen-oxygen gas mixture comprising 33% O2 and 66% H2 of 2000 to 3000 mL per minute by electrolysis of pure water.[

Statistical analysis

SPSS 20.0 was used for statistical analysis. Continuous variables are expressed as median and interquartile range , and classified variables are expressed as frequency and percentage (%). The difference of continuous variables before and after treatment was compared by Mann–Whitney U test, and the difference of classified variables was compared by chi square test and Fisher exact test.

Results

Analysis of clinical baseline indicators

All the patients had a COVID-19 contact history, which was consistent with the epidemiological characteristics. The patients were divided into 2 groups, as those who were subjected to receive additional hydrogen-oxygen therapy (case group, n = 12), and those who were not (control group, n = 12). The clinical baseline indicators are summarized in Table 1. The proportions of male/female in case group and control group are 5/7 and 6/6, and median ages were 51.5 (41.25–59) and 54.5 (40.75–61.5) years, respectively, which are not significantly different between the 2 groups. The clinical types are moderate and severe, which represent the main types of COVID-19. Fever, fatigue, cough, pharyngeal hemorrhage, and tonsil enlargement are the main clinical manifestations. All 24 patients underwent chest CT scan and showed typical imaging features, such as multiple ground glass shadows of the lung. All the patients in case group survived and discharged, while a patient in control group died during hospitalization. These findings indicate that SARS-CoV-2 mainly causes clinical manifestations of inflammation and respiratory infections.

Table 1

Analysis of clinical baseline indicators.

Characteristic	Index	Case	Control	P
Age, median (IQR)—yrs		51.5 (41.25–59)	54.5 (40.75∼61.5)	.644
Age group—n (%)
	<30 yrs	0 (0)	0 (0)	1.000
	30–39 yrs	2 (16.67)	2 (16.67)
	40–49 yrs	3 (25)	3 (25)
	≥50 yrs	7 (58.33)	7 (58.33)
Gender—n (%)
	Female	7 (58.33)	6 (50)	1.000
	Male	5 (41.67)	6 (50)
Smoking—n (%)
	Yes	1 (8.33)	0 (0)	1.000
	No	11 (91.67)	12 (100)
Clinical typing—n (%)
	Mild type	0 (0)	0 (0)	1.000
	Moderate type	9 (75)	10 (83.33)
	Severe type	3 (25)	2 (16.67)
	Critical type	0 (0)	0 (0)
Symptoms—n (%)
	Fever	10 (83.33)	10 (83.33)	1.000
	Chills	2 (16.67)	1 (8.33)	1.000
	Headache	0 (0)	3 (25)	.217
	Fatigue	11 (91.67)	10 (83.33)	1.000
	Shortness of breath	4 (33.33)	1 (8.33)	.317
	Sore throat	4 (33.33)	6 (50)	.680
	Cough	11 (91.67)	10 (83.33)	1.000
	Sputum production	5 (41.67)	3 (25)	.667
	Chest pain	1 (8.33)	0 (0)	1.000
	Diarrhoea	2 (16.67)	1 (8.33)	1.000
Signs of infection—n (%)
	Throat congestion	10 (83.33)	6 (50)	.193
	Tonsil swelling	2 (16.67)	2 (16.67)	1.000
	Enlargement of lymph nodes	0 (0)	0 (0)	1.000
Abnormal Chest CT—n (%)		12 (100)	12 (100)	1.000
Treatment outcome—n (%)
	Death	0 (0)	1 (8.33)	1.000
	Discharge	12 (100)	11 (91.67)

Analysis of related immune inflammation indicators

In this study, we tested and compared the clinical indicators related to immune inflammation in COVID-19 patients, including white blood cell count , neutrophil count , lymphayte count , monocyte count , neutrophil percentage , lymphayte percentage , monocyte percentage, and C-reactive protein (CRP) (Table 2). As shown in Table 2, although the neutrophil percentage and the concentration and abnormal proportion of CRP before therapy have no significant differences between case group and control group, they are significantly lower in case group than in control group after therapy. The results indicate that antivirus, support, and other symptomatic treatments may make limited contribution to reducing inflammation, whereas hydrogen-oxygen therapy can help to suppress inflammatory responses induced by SARS-CoV-2 infection, so as to alleviate the clinical symptoms.

Table 2

Analysis of indicators related to immune inflammation.

	Before therapy			After therapy
Variable	Case	Control	P	Case	Control	P
White blood cell count
Median (IQR)—×109/L	5.27 (4.34–6.14)	4.36 (3.07–7.76)	.488	4.82 (4.19–6.08)	5.49 (4.25–7.03)	.544
> 9.5 × 109/L—n (%)	0 (0)	2 (16.67)	.037	0 (0)	0 (0)
< 3.5 × 109/L—n (%)	0 (0)	3 (25)		0 (0)	0 (0)
Neutrophil count
Median (IQR)—×109/L	3.29 (2.55–4.16)	2.83 (1.63–6.42)	.419	2.55 (2.35–4.32)	3.65 (2.89–5.21)	.126
> 6.3 × 109/L—n (%)	0 (0)	3 (25)	.014	1 (8.33)	0 (0)	1.000
< 1.8 × 109/L—n (%)	0 (0)	3 (25)		0 (0)	0 (0)
Lymphocyte count
Median (IQR)—×109/L	1.22 (1.05–1.7)	0.98 (0.67–1.35)	.133	1.5 (1.14–1.73)	1.19 (0.99–1.51)	.184
> 3.2 × 109/L—n (%)	0 (0)	0 (0)	.414	0 (0)	0 (0)
< 1.1 × 109/L—n (%)	4 (33.33)	7 (58.33)		3 (25)	4 (33.33)
Monocyte count
Median (IQR)—×109/L	0.45 (0.38–0.54)	0.31 (0.16–0.43)	.024	0.36 (0.28–0.48)	0.36 (0.28–0.44)	.643
> 0.6 × 109/L—n (%)	1 (8.33)	1 (8.33)	1.000	0 (0)	0 (0)	1.000
< 0.1 × 109/L—n (%)	0 (0)	1 (8.33)		0 (0)	1 (8.33)
Neutrophil percentage
Median (IQR)—%	63.25 (56.88–68.85)	66.2 (50.58–84.65)	.564	58.1 (52.9–69.75)	66.55 (60.83–76.85)	.030
> 75%—n (%)	2 (16.67)	4 (33.33)	.640	2 (16.67)	4 (33.33)	.640
< 40%—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
Lymphocyte percentage
Median (IQR)—%	24.65 (20.7–32.33)	25.4 (8.55–33.23)	.665	29.45 (20.03–35.95)	21.5 (15.78–29.63)	.112
> 50%—n (%)	0 (0)	1 (8.33)	.1931	0 (0)	0 (0)	.667
< 20%—n (%)	2 (16.67)	5 (41.67)		3 (25)	5 (41.67)
Monocyte percentage
Median (IQR)—%	8.5 (7.38–9.45)	6.35 (4.05–8.2)	.013	6.7 (5.88–8.3)	6.1 (5.43–7.18)	.370
> 10%—n (%)	2 (16.67)	1 (8.33)	.590	1 (8.33)	1 (8.33)	1.000
< 3%—n (%)	0 (0)	2 (16.67)		1 (8.33)	1 (8.33)
CRP
Median (IQR)—mg/L	7.4 (3.45–23.2)	18.15 (4.83–43.33)	.184	2.4 (1.25–6.15)	7.95 (5.05–23.98)	.022
> 4 mg/L—n (%)	9 (75)	10 (83.33)	1.000	3 (25)	10 (83.33)	.012

Analysis of electrolyte-related indexes

In this study, we tested and compared the electrolyte-related indicators of the hospitalized COVID-19 patients, including potassium , sodium , chlorine, total calcium, ionic calcium, CO2-binding capacity, phosphorus , and magnesium (Table 3). The results show that the levels of potassium, sodium, chlorine, phosphorus, total calcium, and CO2-binding capacity of some patients in the 2 groups were abnormal before therapy, while both the routine therapy and hydrogen-oxygen therapy greatly improved electrolyte disturbances. Moreover, there are no significant differences in these indexes between case group and control group after therapy. These findings suggest that hydrogen-oxygen, antivirus, support, and symptomatic therapy may maintain the electrolyte balance and improve respiratory acidosis to a certain extent.

Table 3

Analysis of electrolyte-related indexes.

	Before therapy			After therapy
Variable	Case	Control	P	Case	Control	P
Potassium
Median (IQR)—mmol/L	4.09 (3.79–4.25)	3.66 (3.55–4.14)	.166	4.32 (4–4.47)	4.49 (4.25–4.72)	.166
> 5.5 mmol/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)
< 3.5 mmol/L—n (%)	2 (16.67)	2 (16.67)		0 (0)	0 (0)
Sodium
Median (IQR)—mmol/L	139.36 (138.9–140.48)	139.45 (136.68–140.65)	.419	139.2 (138.57–140.13)	138.72 (138.26–140.01)	.624
> 147 mmol/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)
< 135 mmol/L—n (%)	0 (0)	1 (8.33)		0 (0)	0 (0)
Chlorine
Median (IQR)—mmol/L	99.17 (94.46–101.19)	103.35 (98.05–105.43)	.018	101.73 (100.96–102.4)	100.54 (95.61–101.41)	.053
> 108 mmol/L—n (%)	0 (0)	0 (0)	.640	0 (0)	0 (0)	.093
< 96 mmol/L—n (%)	4 (33.33)	2 (16.67)		0 (0)	4 (33.33)
Total calcium
Median (IQR)—mmol/L	2.31 (2.25–2.36)	2.16 (2.08–2.23)	.004	2.38 (2.37–2.41)	2.37 (2.26–2.4)	.201
> 2.8 mmol/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)	1.000
< 2 mmol/L—n (%)	0 (0)	1 (8.33)		0 (0)	1 (8.33)
Ionic calcium
Median (IQR)—mmol/L	1.15 (1.13–1.18)	1.08 (1.04–1.12)	.005	1.19 (1.17–1.2)	1.18 (1.13–1.21)	.502
> 1.4 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)	1.000
< 1 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	1 (8.33)
CO2-binding capacity
Median (IQR)—mmol/L	27.3 (26.65–30.33)	23.9 (22.35–25.68)	.002	25.2 (24.1–28.8)	27.15 (23.95–30.55)	.506
> 29 mmol/L—n (%)	3 (25)	4 (33.33)	1.000	2 (16.67)	4 (33.33)	.371
< 23 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	1 (8.33)
Phosphorus
Median (IQR)—mmol/L	1.05 (0.92–1.17)	1.02 (0.88–1.14)	.623	1.21 (1.15–1.3)	1.16 (0.98–1.26)	.260
> 1.6 mmol/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)	1.000
< 0.8 mmol/L—n (%)	0 (0)	1 (8.33)		0 (0)	1 (8.33)
Magnesium
Median (IQR)—mmol/L	0.99 (0.96–1.02)	0.96 (0.91–1.02)	.401	0.99 (0.92–1.04)	0.98 (0.89–1.08)	.795
> 1.25 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
< 0.65 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)

Analysis of myocardial enzymes

In this study, we analyzed the related indexes of myocardial enzymes, such as lactate dehydrogenase (LDH), creatine kinase (CK), and CK isoenzymes (CK-MB) in 24 hospitalized COVID-19 patients (Table 4). Before therapy, abnormal changes of LDH, CK, and CK-MB levels were observed in the 2 groups. Except the LDH levels of 2 patients in control group, all the indexes of myocardial enzymes returned to normal range after therapy; however, there are no significant differences in the changes of LDH between case group and control group. The results imply that hydrogen-oxygen therapy may have certain protective effect on the heart of COVID-19 patients.

Table 4

Analysis of myocardial enzyme profile.

				After therapy
Variable	Case	Control	P	Case	Control	P
LDH
Median (IQR)—U/L	220.15 (167.88–248.08)	149.45 (121.08–325.2)	.094	183.7 (135.08–196.38)	172.9 (141.45–227.68)	.707
> 240 U/L—n (%)	3 (25)	3 (25)	1.000	0 (0)	2 (16.67)	.331
< 100 U/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
CK
Median (IQR)—U/L	48.85 (24.45–108.8)	53.8 (44.02–153.05	.273	42.1 (26.53–57.6)	44.4 (40.78–65.73)	.544
> 194 U/L—n (%)	0 (0)	2 (16.67)	.331	0 (0)	0 (0)
< 24 U/L—n (%)	3 (25)	1 (8.33)		0 (0)	0 (0)
CK-MB
Median (IQR)—U/L	8.2 (6.53–9.9)	10.2 (8.88–16.2)	.073	9.15 (7.7–10.65)	7.5 (5.23–12.28)	.452
> 25 U/L—n (%)	0 (0)	1 (8.33)	1.000	0 (0)	0 (0)

Analysis of liver function-related indexes

In this study, we analyzed the liver function-related indexes of the hospitalized COVID-19 patients, such as serum direct bilirubin (DBil), indirect bilirubin (IBil), aspartate aminotransferase (ALT), alanine aminotransferase (AST), AST/ALT, Glutamyl Transferase (GGT), alkaline phosphatase (ALP), total protein (TP), Albumin (Alb), Globulin (Glo), Albumin/globulin, and total bile acid (Table 5). Before therapy, the levels of DBil, ALT, AST, GGT, ALP, TP, Alb, Glo, and Albumin/globulin exhibited abnormal changes in the COVID-19 patients. Although the changes of liver function indexes are not significant between case group and control group after therapy, the control group seemed to be worse than those before therapy. These results suggest that hydrogen-oxygen therapy has no hepatotoxic effect on patients with COVID-19.

Table 5

Analysis of indexes related to liver function.

	Before therapy			After therapy
Variable	Case	Control	P	Case	Control	P
DBil
Median (IQR)—μmol/L	3.85 (2.83–6.28)	3.35 (2.58–4.28)	.470	3.3 (2.55–4.7)	3.6 (2.8–5.45)	.506
> 6.8 μmol/L—n (%)	2 (16.67)	1 (8.33)	1.000	0 (0)	2 (16.67)	.478
IBil
Median (IQR)—μmol/L	6.75 (5.1–10.68)	4.75 (3.97–7.1)	.046	6.2 (4.18–9.15)	6.6 (6.13–8.78)	.583
> 20 μmol/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
ALT
Median (IQR)—U/L	28.65 (19.4–36.33)	23.05 (20.4–36.45)	.603	22.4 (18–36.68)	32.7 (18–48.83)	.507
> 45 U/L—n (%)	1 (8.33)	1 (8.33)	1.000	2 (16.67)	3 (25)	1.000
AST
Median (IQR)—U/L	24.15 (19.75–28.95)	31.15 (22.73–45.6)	.094	20.3 (17.23–27.13)	25.2 (18.85–36.33)	.141
> 45 U/L—n (%)	0 (0)	3 (25)	.217	0 (0)	1 (8.33)	1.000
AST/ALT
Median (IQR)	0.88 (0.66–1.07)	1.5 (0.96–1.75)	.008	0.84 (0.56–1.13)	0.98 (0.66–1.21)	.665
> 2.5—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
GGT
Median (IQR)—U/L	35.05 (24.28–61.78)	32.75 (15.73–65.78)	.564	41.05 (34.18–67.65)	49.35 (29.38∼71.23)	.977
> 50 U/L—n (%)	4 (33.33)	3 (25)	1.000	4 (33.33)	6 (50)	.680
ALP
Median (IQR)—U/L	61.65 (51.73–75.35)	56.7 (49.98–75)	.603	62.2 (52.78–71.75)	65.3 (45.05–76.85)	.707
> 128 U/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)	1.000
< 40 U/L—n (%)	0 (0)	1 (8.33)		0 (0)	1 (8.33)
Total protein
Median (IQR)—g/L	65.95 (62.63–74)	65.7 (61.05–70.53)	.885	68.4 (63.5–70.48)	65.2 (59.8–68.78)	.214
> 85 g/L—n (%)	0 (0)	0 (0)	1.000	0 (0)	0 (0)	.217
< 60 g/L—n (%)	1 (8.33)	1 (8.33)		0 (0)	3 (25)
Albumin
Median (IQR)—g/L	40.45 (35.7–44.45)	41.25 (38.55–44.6)	.623	42.15 (40.03–44.33)	39.85 (33.83–40.78)	.060
> 53 g/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)	.400
< 40 g/L—n (%)	6 (50)	5 (41.67)		3 (25)	6 (50)
Globulin
Median (IQR)—g/L	25.6 (23.5–27.78)	23.25 (21.33–31.13)	.225	25.55 (22.85–28.3)	25 (22.68–32.85)	.885
> 32 g/L—n (%)	2 (16.67)	2 (16.67)	1.000	2 (16.67)	3 (25)	1.000
< 20 g/L—n (%)	0 (0)	0 (0)		1 (8.33)	1 (8.33)
Albumin/globulin
Median (IQR)	1.6 (1.3–1.78)	1.75 (1.35–1.98)	.401	1.5 (1.43–1.9)	1.65 (1.2–1.7)	.621
> 2.5—n (%)	0 (0)	0 (0)		1 (8.33)	0 (0)	1.000
< 1—n (%)	1 (8.33)	0 (0)		0 (0)	1 (8.33)
Total bile acid
Median (IQR)— μmol/L	2.65 (1.63–3.6)	2.9 (1.47–4.03)	.862	2.45 (1.85–4.15)	2.1 (1.8–3.73)	.563
>12 μmol/L— n (%)	0 (0)	0 (0)		0 (0)	0 (0)

Analysis of renal function-related indexes

In this study, we compared and analyzed the renal function related indexes of patients with COVID-19, including urea (Ure), creatinine , uric acid , and cystatin C (CysC) (Table 6). The results showed that the levels of Ure and CysC were abnormal in some patients of 2 groups before therapy; however, there are no significant differences in all the indexes between case group and control group after therapy, indicating that the renal function of patients with COVID-19 would not be affected by hydrogen-oxygen therapy.

Table 6

Analysis of indexes related to renal function.

	Before therapy			After therapy
Variable	Case	Control	P	Case	Control	P
Urea
Median (IQR)—mmol/L	3.13 (1.98–4.19)	3.39 (2.47–4.1)	.371	3.7 (2.4–4.68)	4.06 (3.04–6.99)	.175
> 7.14 mmol/L—n (%)	0 (0)	1 (8.33)	1.000	0 (0)	3 (25)	.093
< 1.43 mmol/L—n (%)	0 (0)	0 (0)		0 (0)	1 (8.33)
Creatinine
Median (IQR)—μmol/L	62.1 (47.45–73.25)	68.9 (61.38–89)	.106	72.5 (60.68–78.53)	62.7 (56.9–76.08)	.402
> 97.2 μmol/L—n (%)	0 (0)	0 (0)		1 (8.33)	0 (0)	1.000
< 35.4 μmol/L—n (%)	0 (0)	0 (0)		0 (0)	0 (0)
Uric acid
Median (IQR)— μmol/L	285.8 (244.83–334.25)	232.45 (218.53–310.63)	.356	271.1 (235.9–366.18)	302.05 (198.13–338.48)	.840
> 430 μmol/L—n (%)	0 (0)	0 (0)		0 (0)	1 (8.33)	1.000
Cystatin C
Median (IQR)— mg/L	0.62 (0.54–0.67)	0.62 (0.6–0.77)	.311	0.65 (0.58–0.72)	0.65 (0.6–0.8)	.707
> 1.05 mg/L—n (%)	0 (0)	0 (0)	.590	0 (0)	0 (0)	1.000
< 0.55 mg/L—n (%)	3 (25)	1 (8.33)		1 (8.33)	1 (8.33)

Discussion

COVID-19 has become a global pandemic infectious disease. Due to the lack of specific antiviral drugs, supportive therapy is used as the main treatment method.[ Clinical studies have shown that the death of COVID-19 patients is characterized by typical pathological changes of lung parenchyma, and eventually develop progressive hypoxemia, lactatemia, acute respiratory distress syndrome , and respiratory failure.[ Hyperbaric oxygen therapy and normobaric oxygen therapy such as nasal duct and mask oxygen inhalation, noninvasive and invasive ventilation, and extracorporeal membrane oxygen have been widely used in the treatment of COVID-19; nonetheless, the mortality of severe patients remains high,[ and even more than 60%.[ This suggests that inhalation of excessive oxygen (O2) may contribute to the detriment of survival in COVID-19 patients. Therefore, it is urgent to explore approaches for elimination of oxygen toxicity.

Hyperoxia, also called excess oxygen, increases formation of reactive oxygen species (ROS), and hence causes multiple-organ toxicity (e.g., central nervous, pulmonary, cardiac, and ocular toxicity).[ In contrast to O2, H2 is reductive, thereby it can selectively scavenge ROS (highly reactive hydroxyl radical, but not super oxide, peroxide, or nitric oxide) and has no serious side effects.[ Thus, we assumed that inhaled O2 added with a certain concentration of H2 may effectively protect organ functions against hyperoxia-induced toxicity. Hydrogen-oxygen nebulizers can provide hydrogen-oxygen mixed gas diffusing into the human body, reduce the flow resistance of gas in the bronchial tree, improve the utilization rate of H2 and O2, so as to provide auxiliary treatment for dyspnea and hypoxemia. In this study, we retrospectively observed the curative effect of hydrogen-oxygen therapy in 12 patients with moderate and severe COVID-19. The results of statistical analysis show marked improvement in immune-inflammation after treatment with hydrogen-oxygen mixture. More importantly, all the patients recovered and finally discharged. These findings suggest that hydrogen-oxygen therapy is conducive to alleviate inflammation, ameliorate dyspnea, and promote survival.

Recent studies suggest that oxidative stress is associated with RNA virus infection and pathogenesis including viral replication, immune response, apoptosis, inflammation, and organ dysfunction; thereby, it can enhance the severity of COVID-19.[ It has been demonstrated that SARS-CoV-2 can substantially promote mononuclear phagocyte system and neutrophil granulocytes to produce ROS which increase the formation of neutrophil extracellular traps , suppress the adaptive immune responses, and cause organ damage.[ The systemic hypoxia induced by persistent hypoxemia should be treated with extra oxygen supply. However, oxygen therapy that may cause oxidative stress often facilitates COVID-19 patients to develop respiratory distress.[ We herein observed significant decrease of the neutrophil percentage and the concentration and abnormal proportion of CRP in COVID-19 patients who had received additional hydrogen-oxygen therapy (Table 2), suggesting that anti-oxidation, anti-virus, and anti-inflammation are probably involved in the underlying mechanism of this treatment method.

In the present study, we observed the abnormal elevations of LDH, CK, CK-MB, DBil, ALT, AST, GGT, Glo, and Ure and the decreases of TP, Alb, and CysC in a part of the patients before therapy by examining myocardial enzyme profile and functions of liver and kidney (Tables 4–6), suggesting that SARS-CoV-2 can damage heart, liver, and kidney. In addition, we found that the antivial and symptomatic therapy seemed to further impair liver function (Table 5), indicating that the antivial drugs may possess hepatotoxicity. Japanese scholars Katherine and Hayashida have demonstrated that breathing 2% H2 can lighten liver and myocardial ischemia-reperfusion injury in animal models.[ However, we found that hydrogen-oxygen therapy suppressed the elevations of LDH and DBil and the decreases of TP and Alb with no statistical significance (Tables 4 and 5), which may be due to the small number of cases.

Conclusion

To our knowledge, this study is the first to demonstrate the curative effect of hydrogen-oxygen therapy on COVID-19; however, the sample size is the primary limitation. A prospective multicenter clinical trial is needed to clarify whether hydrogen-oxygen therapy is superior to hyperbaric oxygen therapy and normobaric oxygen therapy in COVID-19 treatment.

Acknowledgments

The authors thank the patients and staff from all the units who participated in the study, and thank Xiaorong Shui, Dafeng Chen, and Qing He for their contributions to this study.

Author contributions

PL and YD performed the experiments and collected the data. YH, DC, QH, and ZH organize and analyzed the data. WL and YH wrote the manuscript. SH and WL designed this study.

Data curation: Yuan He, Dafeng Chen, Qing He, Zufeng Huang.

Funding acquisition: Shian Huang.

Investigation: Peng Luo, Yuanfang Ding, Shian Huang, Wei Lei.

Methodology: Peng Luo, Yuanfang Ding, Yuan He.

Project administration: Wei Lei.

Resources: Peng Luo, Yuanfang Ding.

Validation: Shian Huang.

Writing – original draft: Wei Lei.