Anti-inflammatory and antioxidant activity of high concentrations of hydrogen in the lung diseases: a systematic review and meta-analysis

As a small molecule, hydrogen is colorless, odorless and lightest. Many studies conducted that hydrogen can protect almost every organ, including the brain, heart muscle, liver, small intestine, and lungs. To verify whether high concentrations of hydrogen (HCH) has anti-inflammatory and antioxidant activities on respiratory system, we product a systematic review and meta-analysis. We investigated MEDLINE-PubMed, Cochrane Library, ScienceDirect, Wiley and SpringerLink database and selected in vivo studies related to the anti-inflammatory or antioxidant effects of HCH in the lung diseases which were published until September 2023. We firstly identified 437 studies and only 12 met the inclusion criteria. They all conducted in rodents. The results showed that HCH had a positive effect on the reduction of tumor necrosis factor alpha (TNF-α), interleukin (IL)-1β, IL-4, IL-8, malondialdehyde (MDA), superoxide dismutase (SOD) and reactive oxygen species (ROS); but there is no effect on IL-6, we speculated that may contribute to the test results for different body fluids and at different points in time. This meta-analysis discovered the protective effects on inflammation and oxidative stress, but whether there exists more effects on reduction of inflammatory and oxidant mediators needs to be further elucidated.


Background
By now, the incidence of respiratory diseases is increasing, and its mortality rate is among the top three in the world.It also imposes a huge economic burden.In addition, it has an emotional impact on patients (1).. Therefore, there needs a more efficient and economical method to save patients' survival and quality of life.Many researchers believe that pulmonary inhalation may be a more direct and effective way with fewer side effects.Many similar studies are under way.Hydrogen, for example, is a treatment that is inhaled directly into lungs.
Hydrogen (H 2 ), a diatomic gas composed of two hydrogen atoms connected by covalent bonds, is produced by the intestinal bacteria of mammals; H 2 is colorless and odorless and is a stable neutral molecule (2).In 2007, Ohsawa et al (3) reported that H 2 can react with cytotoxic oxygen free radicals by reacting with hydroxyl free radicals (•OH) in cultured cells.H 2 does not react with •O 2-, H 2 O 2 or NO.Due to its potential ability to anti oxidative stress, inflammation, and apoptosis, H 2 is emerging as the fourth gas signaling molecule in the body (4).Generally, hydrogen concentrations between 4% and 75% will not increase, and this paper defines hydrogen concentrations above 4% as highconcentration hydrogen (HCH).A systematic review by Yuan et al. (5) reported its potential protective effects on ischemia/ reperfusion injury in multiple organs, neurodegenerative diseases, bone and joint diseases, and respiratory diseases.
The commonly used hydrogen administration methods include direct inhalation of hydrogen, injection of hydrogen-rich water and oral hydrogen-rich water (6).This paper mainly explored the therapeutic effect of hydrogen inhalation on respiratory diseases.In 1975, American scholar Dole et al. (7) reported in Science that inhaling hydrogen at 8 atmospheres for 14 consecutive days could significantly reduce the size of skin cancer tumors in mice; this was the first study in human history to determine the medical effect of hydrogen.In 2007, Wood et al (8) evaluated hydrogen as a cytoprotective therapy for ischemia-reperfusion injury and stroke, calling it a selective antioxidant with explosive potential, and this effect has also been confirmed in human experiments (9).At first, most experiments explored the therapeutic effects of low concentrations of hydrogen, but considering the actual concentration of hydrogen inhaled in the body, a higher concentration of hydrogen was derived.
Clinically, Chen et al. (10) reported that inhaling 67% hydrogen can alleviate the disease progression of non-small cell lung cancer; Zheng et al. (11) found that hydrogen therapy can treat acute episodes of chronic obstructive pulmonary disease (COPD); Akagi et al. (12) found that hydrogen can improve the prognosis of advanced colorectal cancer patients; Zeng et al. (13) reported that in the treatment of COVID-19, a mixture of hydrogen and oxygen can improve patients' percutaneous arterial oxygen saturation (SpO 2 ) and shorten the length of hospital stay.Some animal experiments have shown that a high concentration of hydrogen can reduce the secretion of inflammatory factors, possibly through a variety of signaling pathways, such as nuclear factor-kappa B (NF-kb), and can reduce the content of reactive oxygen species (ROS) and some oxidation products.
These studies verified the therapeutic effects of HCH.In order to further evaluate its anti-inflammatory and antioxidant capacity in respiratory diseases, we demonstrated this through this systematic review and meta-analysis.

Methods
We conducted this study following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement ( 14) and Cochrane Manual.

Eligibility criteria
Our criteria for inclusion are: (a) the experimental model was an animal model with lung disease; (b) the intervention was treatment inhalation with high concentration hydrogen alone; (c) the results are an indicator of anti-inflammatory or antioxidant outcomes in the treatment of lung disease; and (d) the type of study was experimental.
All the retrieved titles, abstracts, and full texts were read and screened independently by at least two researchers.If a disagreement arises, it is discussed with reference to the inclusion exclusion criteria.The inclusion criteria were as follows: animal studies; suffers from respiratory problems; high concentration hydrogen inhalation was used alone; and anti-inflammatory or antioxidant outcome measures were used.The exclusion criteria were as follows: articles that do not meet the inclusion criteria, review articles, meta-analyses, abstracts, conference proceedings, editorials/letters, and case report.

Search strategy
Five databases were used to search for papers that met the criteria of the study: the National Library of Medicine (MEDLINE-PubMed), Cochrane Library, ScienceDirect, Wiley and SpringerLink databases.Different combinations of the following keywords were used: "hydrogen," "respiratory tract diseases," "respiratory system," "lung injury," "pulmonary" and "trachea".
The search strategy is as follows: (hydrogen gas) AND (respiratory system disease or lung disease or pulmonary disease or trachea disease).In addition, we checked the references in the article to make sure there were no potential missing articles.
The databases were searched for studies published until September 2023.This retrieval strategy was used to search for the anti-inflammatory and antioxidant effects of high concentration hydrogen in animal models of respiratory diseases.After reading the retrieved literature, we investigated the relevant references and included the relevant articles in the study.We did not contact the original author when there was data in the article that was not available, nor did we cite data from unpublished articles.

Data collection process and study risk of bias assessment
The data were extracted by one researcher according to Table 1 and examined by another researcher.The data to be extracted were as follows: study design, animal model studied, methodological characteristics of high concentrations of hydrogen, respiratory injury studied, markers evaluated, main results, conclusions.
According to the Cochrane Manual, since we analyzed fewer than 10 articles in each group, we used SYRCLE's risk of bias (RoB) tool for animal studies to assess the study risk of bias.Each article was evaluated by two different researchers, and if there was any disagreement, it was resolved through negotiation.The risk of bias was rated as low, uncertain, or high.The contents include Selection bias, Performance bias, Detection bias, Attrition bias, Reporting bias and Other bias.

Synthesis methods
We used standardized mean difference (SMDs) of 95% confidence intervals (CI) to evaluate the treatment effect.If SMD = 0, it indicates no difference, SMD > 0 indicates more occurrence in the experimental group, and SMD < 0 indicates less occurrence in the experimental group.The mean and standard deviation (SD) of the control and treatment groups are obtained by extracting graphs in the article, and the effect size of the target outcome will be calculated.The negative effect size indicated that HCH could effectively reduce inflammatory mediators and markers of oxidative stress, and the positive effect size indicated that HCH could effectively reduce oxidative stress response for superoxide dismutase (SOD).
We used forest maps to graphically represent the effect size and 95% CI.We used Z test to evaluate the overall effect.If P < 0.05, it indicates that there was a significant difference.We used Chi 2 test to evaluate the heterogeneity of the literature.If I 2 < 50% or P > 0.1, the heterogeneity is small, and the fixed-effect model is used.If I 2 ≥ 50% or P ≤ 0.1, it indicates that there is large heterogeneity in the study, and a random-effects model is used.We performed subgroup analyses based on the markers analyzed.If the results of the study did not include numerical values for the target results, we used the software GetData Graph Digitizer to evaluate their result graph to get an average and SD.We used Review Manager 5.3 (RevMan), 2014) for all of our analyses.

Studies selection
The steps of article retrieval filtering are shown in Figure 1.We identified 437 articles from the five databases and the bibliographies  of relevant articles.By reading the title and abstract of the article, we got 86 articles after eliminating irrelevant articles.After excluding papers not shown in full, duplicates, letters, case studies and those whose themes did not match the criteria of this study, 12 articles remained (Figure 1).The two researchers who screened the articles had a high degree of agreement on inclusion and exclusion (Kappa index >96%).

Included studies characteristics
We selected 12 studies conducted in China.These studies were published between 2018 and 2023 (Table 1).
The levels of inflammatory markers, such as TNF-a, decreased in all the studies in which TNF-a was analyzed (15, 16, 19-21, 23, 26).IL-8 levels decreased in three studies (15,16,21).In three studies, there were no differences in IL-6 compared to that in the control group (22,23,26), but in two other studies (19,20), there was an improvement in this marker.IL-4 decreased in three studies in which it was analyzed (17,24,26), but there were no differences in one study (Wei et al, 2023) (22).Oxidative stress, shown by the MDA levels, was lower in the high-concentration hydrogen group in every study in which it was analyzed (15-17, 23, 26).SOD levels were greater in the high-concentration hydrogen group in every study in which it was (17,21,25).ROS decreased in all the studies in which ROS were analyzed (20, 21).

Data synthesis
In the literature we searched, most of the studies assessed the expression levels of different markers, which made it impossible to conduct a uniform meta-analysis of all the literature, so we conducted a subgroup analysis of the consistent results in some of the literature.Among them, the anti-inflammatory effect of high concentration hydrogen was evaluated using IL-1b, IL-4, IL-8, IL-6 and TNF-a as inflammatory mediators, and the antioxidant effect was evaluated using MDA, SOD and ROS markers.
From Figure 2, we can see the protective effect of high concentration of hydrogen on the reduction in IL-1b (SMD = −2.51,95% CI −3.84 to −1.19, P < 0.005).The I 2 was 78% and P = 0.0001, indicating that there is high heterogeneity in all studies of IL-1b (Figure 2).In order to reduce heterogeneity, we excluded lowquality literature and left two high-quality literature (du et al, 2022; feng et al, 2019) (15,16).I 2 < 50%, the fixed-effect model was used to analyze the results, which showed little difference from the original results and that means good stability.From Figure 3, we can see the positive effect of high concentration of hydrogen on the reduction in IL-8 (SMD = −1.95,95% CI −3.86 to −0.04, P = 0.05).The I 2 was 84% and P = 0.002, indicating that there is high heterogeneity in all studies of IL-8 (Figure 3).In order to reduce the heterogeneity, the group with the smallest sample size was excluded (15), and the heterogeneity was reduced to 52%, which still had a significant difference.However, the heterogeneity of retained high-quality literature is still high, and the results are not significant and the results are poor in stability.From Figure 4, we can see the protective effect of HCH on the reduction in TNF-a (SMD = −2.98,95% CI −4.25 to −1.71, P < 0.005).The I 2 = 77% and P = 0.0002, indicating that there is high heterogeneity in all studies of TNF-a (Figure 4).In order to reduce heterogeneity, we retained highquality literature for analysis (15,16,26), and the heterogeneity became smaller, I 2 < 50%, and the results still had significant differences and they have good stability.From Figure 5, we can see the positive effect on the reduction in IL-4 (SMD = −1.87,95% CI −3.14 to -0.6, P < 0.05.The I 2 = 67% and P = 0.03, indicating that there is high heterogeneity in all studies of IL-4 (Figure 5).In order to reduce heterogeneity, studies with a small sample size were eliminated (22,24), and the heterogeneity became smaller with I 2 < 50%, indicating little difference in results and have good stability.From Figure 6, we can see there was no effect on IL-6 (SMD = −0.71,95% CI −2.14 to 0.72, P = 0.33).The I 2 = 83% and P = 0.0001, indicating that there is high heterogeneity in all studies of IL-6 (Figure 6).To reduce heterogeneity, we assessed the quality of the literature, but there was only one high-quality literature (zhang et al, 2018) (26), and the heterogeneity was high regardless of the group, and there was no significant difference in the results.
From Figure 7, we can see the protective effect of HCH on SOD (SMD = 3.22, 95% CI 0.43 to 6.01, P < 0.05).The I 2 = 87% and P = 0.0004, indicating that there is high heterogeneity in all studies of SOD (Figure 7).In order to reduce the heterogeneity, the minimum SMD was removed for analysis (huang et al, 2019) (17), and the heterogeneity was reduced with I 2 < 50%.The fixed-effect model was used for analysis, and the results showed no significant difference and have good stability.From Figure 8, we can see the positive effect on the reduction in ROS (SMD = −2.71,95% CI −4.84 to −0.59, P < 0.05).The I 2 = 82% and P = 0.01, indicating that there is high heterogeneity in all studies of ROS (Figure 8).However, there were only two literatures in this group, which could not narrow the heterogeneity for subgroup analysis.From Figure 9, we can see the positive effect on the reduction in MDA (SMD = −1.65,95% CI −2.60 to -0.71, P < 0.05).The I 2 = 52% and P = 0.10, indicating that there is high heterogeneity in all studies of MDA (Figure 9).In order to reduce the heterogeneity, we conducted subgroup analysis of the experimental animal with mouse, the group whose experimental animals were rats was excluded (feng et al, 2019) and the results showed that the heterogeneity was reduced ( 16), I 2 < 50%, but the results were not different and the stability was good.

Risk of bias
Table 2 summarizes the risk of bias of the 12 studies based on the Systematic Review Centre for Laboratory Animal Experimentation (SYRCLES) risk of bias tool.In Sequence generation, only one study had a low risk (25), they described using a random number table in the literature, while the other studies had unclear risk, there is no evidence how random sequence produced.For baseline characteristics, two studies had low risk (17,21), they conducted baseline measurement in the article, while others did not describe and had unknown risk.The risk of allocation concealment in all studies is unknown, they all not specify whether there is allocation concealment.Almost all studies showed a low risk in performance bias, they described the same feeding environment and the same administration conditions.For the random outcome assessment, there were five studies with high risk (15,16,21,24,26), because that did not describe using a random number table to choose experimental animals and the remaining risks were unclear.Meanwhile, six studies showed low risk in detection bias of blinding (15-17, 21, 24, 26), because almost all study used all animal's results in the outcome that indicate there is no detection bias.There were six studies with low risk in incomplete outcome data (15-17, 21, 24, 26), because all animals were absorbed in the outcome.The rest with high risk (18-20, 22, 23, 25), because they did not describe how to deal with missing data.In reporting bias and other bias, all studies showed low risk.All the data described in method has been reported in results and there is no drug sharing and undue influence from funders.So, they all in low risk.

Discussion
According to Matei et al. ( 27), the therapeutic potential of hydrogen has received much attention, and researchers have reported that hydrogen has a beneficial effect on a variety of diseases, including lung diseases such as COPD and ALI.
Other studies have shown that HCH has many pharmacological properties, such as antioxidant and anti-inflammatory effects.The anti-inflammatory effect of HCH may be mediated by the regulation of NF-kB (28).The forest plot (Figure 4) shows that HCH has a positive effect on reducing TNF-a, and all of the analyzed studies included showed that HCH was able to reduce this inflammatory mediator.According to Gardam (29), TNF-a is a key mediator of the activation and recruitment of inflammatory cells, including polymorphonuclear neutrophils (PMNs) and macrophages.In addition, it can also induce the release of proinflammatory markers and oxidative and nitrosation stress in the lung endothelium (30,31).According to Carvalho (1), the primary action of IL-8 is to stimulate the migration of immune system Meta-analysis of IL-1b differences-the HCH group versus the control group.26) did not discovered significantly change in the level of IL-6, which make it no significant difference in the result of metaanalysis.We hypothesize that the non-significant difference in results may be due to the following reasons: first, the sources of IL-6 measured in literature are different.When detecting IL-6 in BALF in zhang et al (26), there is no significant difference in hydrogen group, but there is a significant difference in serum, but the data processing software cannot obtain this result.Therefore, only IL-6 levels in BALF were analyzed.In addition, yin et al (23) analyzed IL-6 levels in different time periods, each time point showed different therapeutic effects, but in this study, we only analyzed one of the time points, so IL-6 levels are constantly changing in the course of disease and treatment.This time point we chose is not representative of the therapeutic level of hydrogen in IL-6 over the course of treatment.
De Carvalho et al. (37) reported that smoke inhalation can cause lung and systemic lesions, mainly involving inflammatory processes and oxidative stress, in which the oxidative stress mediators include MDA and so on.MDA is an important marker of oxidative stress, and HCH can significantly reduce its production (Figure 9).In addition, in a population study, it was found that HCH can significantly reduce the level of MDA in lung disease (38).This study confirms the conclusion that HCH can downregulate the level of oxidative stress in the body in lung disease.The quantification of MDA in biological systems is an important parameter for evaluating cellular oxidative stress and is used to estimate lipid peroxidation in the lung (37, 39).Many studies with hydrogen, which uses this marker to analyze the progression of the inflammatory process, have shown that it has positive effects (40,41).
ROS include free radicals, such as •OH, superoxide anion radicals (O 2 • − ), and nonfree radical species, such as singlet oxygen ( 1 O 2 ) and H 2 O 2 (42).They are generated inside the body by aerobic organisms as a byproduct of energy metabolism through oxidative phosphorylation (43).Normally, there are antioxidant defense systems in cells that protect biological systems from free radical toxicity, such as SOD, catalase (CAT), glutathione peroxidase (GSH-Px), and heme oxygenase-1 (HO-1) (44,45).The effects of HCH on ROS and SOD were studied in this paper.Positive effects were observed in the included studies.
Hancock et al. (46) states that hydrogen, a well-known antioxidant, has possible positive effects on lung diseases.Its oxidative stress-reducing parameters have been widely studied in several pathologies (47)(48)(49).
However, this study has certain limitations.All the included studies were from China, and the application of high concentrations of hydrogen needs to be confirmed by more researchers.In addition, for cases with heterogeneous sources or no significant difference, we analyzed that the therapeutic effect of hydrogen in the indicators of inflammation and oxidative stress changes with time, so there should be studies to explore which stage hydrogen plays the most important role in the occurrence of inflammatory factors or oxidative stress.Or explore whether hydrogen mainly works by acting directly on the lungs or after entering the blood.Finally, most of the published clinical studies on hydrogen in the respiratory system are on HCH, and the detection indicators are limited to the relief of clinical symptoms.More in-depth clinical studies need to be carried out.

Conclusion
Our results suggest that high concentrations of hydrogen have anti-inflammatory and antioxidant effects in certain inflammatory or oxidative stress mediators.However, at present, there are problems such as small sample size of animal studies or small number of human experiments.A more targeted experimental design would make it possible to more clearly elucidate the relationship between high concentrations of hydrogen and Meta-analysis of malondialdehyde (MDA) differences-in the HCH group versus the control group.
mediators of inflammation and oxidative stress.For now, more high-quality studies are needed to validate these findings.Whether there are more effects on reducing inflammation and oxidation mediators remains to be further elucidated.

FIGURE 1
FIGURE 1Flow diagram of the literature search and screening process.

FIGURE 3
FIGURE 3Meta-analysis of IL-8 differences-the HCH group versus the control group.

TABLE 1
Description of the main aspects of the studies included in the systematic review.