The impact of probiotics supplementation on gut microbiota after Helicobacter pylori eradication: a multicenter, open-label, randomised trial

Background: Helicobacter pylori (H. pylori) eradication therapy may lead to the perturbation of gut microbiota. The role of probiotics in gut microbiota during eradication therapy is still debated. Design: This was a multicentre, open-label, randomised trial done at seven hospitals in China. 162 patients were enrolled, 79 patients were randomly divided into group A (bismuth quadruple therapy), and 83 patients were randomly subjected into group B (bismuth quadruple therapy supplemented with Medilac-S). Faecal samples were collected before treatment and 2 weeks, 4 weeks, 6 weeks, and 8 weeks after eradication therapy. Gut microbiota was analyzed by 16S rRNA high-throughput sequencing. This trial is complete and registered with Chinese Clinical Trial Registry (Chictr.org.cn, ChiCTR1900022116). Results: The eradication rates of group A and group B were 82.43% and 87.01%, respectively (P>0.05). Compared with baseline, alpha and beta diversity was signicantly altered 2 weeks after eradication in both group A and group B, which was restored at week 8. There were no signicant differences in alpha and beta diversity between the two groups. Bismuth quadruple therapy resulted in enrichment of some detrimental bacteria taxa such as Klebsiella and Streptococcus that were not recovered by week 8. Probiotics supplementation could rapidly restore the taxa levels of Klebsiella and Streptococcus by week 4 after eradication, and increase the benecial taxa of Bacillus and Lactobacillales. Functional analysis revealed that lipopolysaccharide biosynthesis and polymyxin resistance pathways were signicantly enriched after eradication therapy, while probiotics supplementation mainly enriched the cofactors and vitamins metabolism pathways. Several detrimental taxa were identied to be correlated with features of older age, alcohol use and high BMI, while probiotics supplementation could effectively restore the adverse impact in patients with these characteristics.


Introduction
H. pylori (H. pylori) is a gram-negative bacterium that infects over half of the world population, and causes various gastric diseases including gastritis, chronic atrophic gastritis, peptic ulcers, gastric mucosa associated lymphoid tissue (MALT) lymphoma, and even gastric cancer [1,2]. Therefore, H. pylori eradication therapy is recommended to reduce the recurrence of peptic ulcer disease and the incidence of gastric cancer [3,4].
Recently, the regimen of H. pylori eradication therapy includes dual therapy, triple therapy, and the bismuth-containing quadruple therapy [5]. Triple therapy remains the standard of care in the published international guidelines of the European Helicobacter and Microbiota Study Group in areas of low clarithromycin resistance [3]. However, the clarithromycin-or metronidazole-based triple therapy for H. pylori infection is no longer recommended in China because of its high antibiotic resistance and low eradication e cacy. Bismuth quadruple therapy is recommended as the rst-line eradication regimen in China [6]. No matter which regimens were used in eradication therapy, several concerns and barriers regarding the widespread use of the antibiotics and proton pump inhibitors (PPIs) were raised [7][8][9][10][11]. H. pylori eradication consists of PPIs and antibiotics can cause the disruption of gut microbiota, which is considered as a major contributing factor in pseudomembranous colitis associated with Clostridium di cile infection, diarrhea, or antibiotic resistance [12]. PPIs can alter gastrointestinal pH which might affect gut microbiota and the survival of enteric pathogens. Moreover, the administration of broadspectrum antibiotics can reduce microbiota diversity, disrupt the microbiota, and enrich the antibioticresistant strains [13]. Eradication therapy may induce the pathogenesis of various disorders through the changes and dysbiosis in gut microbiota. Recently, several studies have assessed the impact of H. pylori eradication on the gut microbiota. Shortly after the H. pylori eradication therapy, the bacterial diversity was signi cantly reduced [8, [14][15][16]. Therefore, studying the effects of eradication therapy on the composition of the gut microbiota and exploring the potential strategy to maintain the microbiota homostasis is extremely important.
Probiotics are kinds of microbes that are bene cial for the host health. Previous studies used the certain probiotics during eradication therapy to decrease side effects, improve compliance, and thereby increase eradication rates [5]. A study conducted in Spain where 209 consecutive patients were prescribed eradication therapy and randomly received probiotics (Lactobacillus plantarum and Pediococcus acidilactici) or matching placebo showed that the eradication rates and side effects were observed to be similar [17]. While some other studies showed that probiotics improved the eradication rate and decreased the incidence of diarrhea, abdominal distension and constipation [18]. Therefore, the role of probiotics in eradication therapy is still debated. Additionally, previous studies mainly focused on the effect of probiotics on the eradication rate and side effects with the standard triple therapy, few studies have analyzed the in uence of probiotics on the gut microbiota on a community-wide scale and the function of the gut microbiome in bismuth quadruple therapy. Especially, the optimal supplementation such as the species, duration, dosage, and the suitable population is extremely limited.
In this study, we aimed to investigate the effects of probiotics supplementation on the homostasis and functional potential of gut microbiota after bismuth quadruple therapy, and further explore the optimal use of probiotics for the suitable people to gain the most bene cial effects of probiotics intake.

Patients and study design
This multicenter, open-label, randomized clinical trial was performed at seven hospitals in China from March 2019 to November 2019. Inclusion criteria: patients aged between 18 and 60 years, at least two positive tests of rapid urease test and 13 C-urea breath test ( 13 C-UBT). Patients with any one of the following criteria were excluded from the study: history of gastrectomy, previous eradication therapy for H. pylori, peptic ulcer or other upper gastrointestinal lesions, gastrointestinal malignant tumor, contraindication or previous allergic reactions to the study drugs, severe concurrent diseases or malignancy, pregnant or lactating women, the use of antiacids or gastric mucosal protective drugs or antibiotics or probiotics in the past two weeks, and patients who could not give informed consent. Written informed consent was obtained from all patients before enrolment, and this trial was approved by the Institutional Review Board of each participating hospital. This trial was registered at Chinese Clinical Trial Registry (Chictr.org.cn, ChiCTR1900022116).
Based on the inclusion and exclusion criteria, a total of 162 patients were enrolled in this study. Eligible patients were randomly assigned to two groups (group A and group B). Patients in group A received 14day bismuth quadruple therapy (BQT) consisting of esomeprazole 20 mg, amoxicillin 1000 mg, furazolidone 100 mg, bismuth potassium citrate 220 mg, all given twice daily. This eradication therapy was given due to the low rates of antibiotic resistance to amoxicillin and furazolidone in southwest China. Patients in group B received the 14-day BQT therapy supplemented with probiotics (Medilac-S; Enterococcus faecium 4.5 × 10 8 and Bacillus subtilis 5.0 × 10 7 , Hanmi, Beijing, China) three times a day for 4 weeks. Gastrointestinal symptoms were assessed at baseline for all patients and on week 2, week 4, week 6, and week 8 after therapy. 13 C-UBT was used to evaluate the H. pylori eradication effect at 6 weeks after completion of treatment.

Fecal sample collection
Fresh stool samples were collected from all patients at baseline (before treatment), at week 2, week 4, week 6, and week 8. Participants were asked to return the faecal specimen to the research assistant in the hospital on the day of sample collection. All stool samples were immediately frozen and stored at -80 °C.
16S rRNA gene ampli cation and sequencing Total DNA from the fecal samples was isolated from the TIANamp Stool DNA Kit (TIANGEN Biotech Co. Ltd., Beijing, China) according to the manufacturer's instructions. DNA concentration was quanti ed using a Nanodrop (Thermo Scienti c, Wilmington, USA), and its integrity was assessed by 1% agarose gel electrophoresis. DNA was stored at -20 °C until use. The V3-V4 hypervariable regions of the 16S rRNA were ampli ed using the following primers: forward Primer:

Sequence analysis
The Quantitative Insights into Microbial Ecology 2 (QIIME2, version 2019. 7) platform in our Ubuntu Linux server was used to process the sequencing data. In the bioinformatics analysis, the fastq format sequence les (spanning the entire 16S rRNA gene V3-V4 region) were loaded into fastqc for basic quality control. Pair-end sequences that passed the basic quality control were then merged and denoised using "qiime dada2 denoise-paired" command of QIIME2 with parameters: p-trim-left-f = 9, p-trim-left-r = 9, ptrunc-len-f = 250, p-trunc-len-r = 250. DADA2 is a pipeline for detecting and correcting (where possible) Illumina amplicon sequence data [19]. DADA2 group unique sequences and referred to as sequence variants (the equivalent of 100% OTUs in QIIME 1) [19]. As implemented in the q2-dada2 plugin, this quality control process will additionally lter any phiX reads (commonly present in marker gene Illumina sequence data) that are identi ed in the sequencing data, and will lter chimeric sequences. The "qiime feature-table summarize" command was used to generate a feature-table listing how many sequences are associated with each sample and with each feature. The lowest sequences count of all the samples were used as the sampling-depth. Then the "qiime feature-table rarefy" command was used to subsample frequencies from all samples so that the sum of frequencies in each sample is equal to sampling-depth [20].
All the alpha diversity and beta diversity results were calculated by "qiime diversity alpha" and "qiime diversity beta" command from the subsampled feature-table (also called rare ed table). The PCoA results were calculated by "qiime diversity pcoa" command and visualized by "qiime emperor plot" command.
A pre-trained Naive Bayes classi er (silva-132-99-nb-classi er.qza) and the "qiime feature-classi er classify-sklearn" command were used to explore the taxonomic composition of the samples. This classi er was trained on the Silva database version 132 [21] with 99% similarity, where the sequences have been trimmed to only include the V3-V4 region of the 16S that was sequenced in this analysis. The Wilcoxon tests were used to evaluate the ecological similarity between and within groups. Furthermore, linear discriminant analysis effect size (LEfSe) was used to select signi cant candidates at the genus level. We compare relative abundance of taxa between the two groups and different time period within each group using a nonparametric Mann-Whitney U test, followed by a Linear Discriminant Analysis (LDA) to estimate the effect size of each microbial feature with differential abundance. It was considered as signi cantly enriched taxa with LDA score greater than 2.0 at a P value < 0.01.

Functional pathway prediction
The 16S rRNA functional prediction by used amplicon sequence variants (ASVs) is performed by PICRUSt2 (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States, version 2). Then ASVs were categorized into Clusters of Orthologous Groups (COG) and into Kyoto Encyclopedia of Genes and Genome (KEGG) orthology (KO). According to the COG database, the descriptive information of each COG and its functional information were parsed from the eggNOG database to obtain the functional abundance spectrum. KO, Pathway and Enzyme (EC) information were obtained according to the KEGG database while the abundance of each functional category was calculated according to OTU abundance.

Clinical correlation analysis
To analyze the correlation between gut microbiota and clinical demographic variables of these selected patients, we employ the non-parametric multivariate analysis of variance (Adonis), an analysis utilizing a matrix of arbitrary squares computed pairwise distances in replace of the covariance matrix that are decomposed into within and between group sums of squares. This analysis was conducted employing the adonis function from the R package vegan. After the signi cant variables were identi ed, the speci c bacteria taxa were identi ed based on the different clinical features using LEfse analysis.

Statistical analyses
Data was presented as mean ± standard deviation (SD). We used the χ2 test or Fisher's exact test for analysis of categorical data and Student's t test or the ANOVA test for analysis of continuous data. Eradication e cacy was performed on an intention-to-treat (ITT) population where patients who dropped out were considered as treatment failures. Secondary per-protocol (PP) analyses were performed which excluded patients lost to follow-up or prematurely withdrew before completion of the study. Differences in relative abundance of bacteria taxa between group A and group B were compared by Wilcoxon rank sum test. All tests were performed using GraphPad Prism

Baseline characteristics
In this study, 162 patients were enrolled, 79 patients were randomly divided into group A, and 83 patients were randomly subjected into group B. 151 patients completed this trial, while 5 patients in group A and 6 patients in group B withdrew from this study (Fig. 1). Finally, a total of 755 stool samples were collected and subjected to analysis. The baseline characteristics of enrolled patients are presented in Table 1. No signi cant differences in age, gender, body mass index (BMI), smoking habits, alcohol consumption, marital status, education and exercise were found between the two groups (all P > 0.05). Before the eradication therapy, there was no signi cant difference in overall gastrointestinal symptoms between the two groups (Supplementary Table S1). Since the diet components had potential impact on the gut microbiota, we then assessed nutrient intake using a Food Frequency Questionnaire (FFQ) in these patients and found that there was no difference between the two groups (Supplementary Table S2). The eradication rates of group A and group B were 82.43% and 87.01% by ITT analysis. By PP analysis, the rates in group A and group B were 84.72% and 89.33%, respectively. No signi cant differences in eradication rates were observed (P > 0.05) ( Table 2).  Alterations in gut microbial diversity after H. pylori eradication and probiotics supplementation We rst evaluated the alpha diversity in each group using four indices chao 1, observed OTUs, Shannon and faith_pd index. The alpha diversity indices were signi cantly decreased 2 weeks after treatment in both groups (P<0.01), which lasted for up to 4 weeks after treatment (P<0.05). There was a trend for restoration of microbiota with time. The alpha diversity almost returned to the baseline at week 8 ( Figure  2A, Supplementary Figure S1). As for the comparison between the two groups, there were no signi cant differences in alpha diversity at baseline, week 2, week 4, week 6 and week 8 (P>0.01; Figure 2B and C, Supplementary Figure S2).
In order to examine the variability of microbial community between the two groups, we calculated beta diversity using PCoA on weighted Uniface distance. There were no signi cant differences in the beta diversity between the two groups before treatment (P=0.652; Figure 2D). However, there were signi cant differences in the beta diversity at week 2, week 4, week 6 and week 8 after treatment compared to the baseline based on all patients or each group alone ( Figure 2E and F). There were no signi cant differences in beta diversity between the two groups at each time period ( Figure 2F).

Changes in microbiota taxa after H. pylori eradication and probiotics supplementation
In order to identify the pro les of gut microbiome changes, we examined the microbiota taxonomic composition and relative abundance in the two groups at different taxonomic levels. At the phylum level, the relative abundance of Proteobacteria was signi cantly increased, while the abundance of Firmicutes, Bacterioidetes, Verrucomicrobia, and Actinobacteria was signi cantly decreased 2 weeks after treatment in both groups ( Figure 3A). No signi cant difference in phyla was observed at week 4, week 6, and week 8 compared to the baseline ( Figure 3A). Meanwhile, there was a decrease in the Bacterioidets:Firmicutes (B:F) ratio at week 2 after treatment which returned to the baseline from week 4 to week 8 ( Figure 3B). At genus level, we observed an increased abundance of Clostridium, Klebsiella, Streptococcus, and Veillonella, while a decrease of Bacteroides, Faecalibacterium, Roseburia, Lachnospira, Phascolarctobacterium, Megamonas, Oscillospira, and Ruminococcus at 2 weeks after treatment compared to the baseline in both groups ( Figure 3C and D).
We further used LEfSe to speci cally identify the bacterial taxa in each group at different periods after eradication. In group A, several bacterial taxa were differentially abundant compared to the baseline with linear discriminant analysis (LDA) score>2 and P < 0.01. Klebsiella, Streptococcus, Veillonella, Fusobacterium, Morganella, and Prevotella were signi cantly enriched, while Bacteroides, Faecalibacterium, Roseburia, Lachnospira, Phascolarctobacterium, Bi dobacterium, and Butyricimonas were markedly decreased at week 2 (Supplementary Figure S3A). To specify the distinct microbial taxa between the two groups at each time point, we further performed LEfSe to compare bacterial abundances. We observed that Enterococcus, Citrobacter, and Oscillospira were signi cantly enriched in group B, while Dialister, Anaerotruncus, and Megasphaera were mainly enriched in group A at week 2 ( Figure 4A). At week 4, Enterococcus, Bacillus, and Lactobacillales were enriched in group B ( Figure 4B), suggesting the successful colonization of probiotics after its supplementation. By week 6 and week 8, the enriched abundance of Enterococcus and Bacillus disappeared ( Figure 4C and D). The above results suggested that H. pylori eradication could signi cantly disturb the composition of gut microbiota, enrich some detrimental bacteria taxa such as Klebsiella and Streptococcus, while decrease some bene cial taxa like Faecalibacterium, Roseburia, Lachnospira, Phascolarctobacterium, Bi dobacterium, and Butyricimonas. The effective colonization of probiotics could be observed during its supplementation, and the probiotics supplementation might rapidly decrease the enrichment of some detrimental bacteria taxa.
Predictive functional pathways of microbial community after H. pylori eradication and probiotics supplementation To investigate the potential role of gut microbiome, we tried to identify the functional variations in the microbial communities using PICRUSt analysis with KEGG database to predict microbiota associated functional pathways. We observed differential predicted functions between samples before and after H. pylori treatment. At week 2 in group A, pathways involved in starch degradation, pyruvate fermentation, glycolysis, glucarate degradation were decreased, whereas those involved in fatty acid oxidation, sucrose degradation, formaldehyde assimilation were increased at week 2 (Supplementary Figure S4A). Consistent with the tendency of microbial changes, we observed some changes of the predicted pathways were also transient, while some other changes such as an increase of sucrose degradation and fatty acid oxidation as well as the decrease of starch degradation and glycolysis were longer lasting until week 8 (Supplementary Figure S4A-D). In group B, pathways involved in propanediol degradation, fatty acid oxidation, pyruvate_dehydrogenase, hexitol_fermentation sucrose_degradation, methylphosphonate_degradation were increased, whereas those involved in glycogen_degradation, peptidoglycan_biosynthesis, L_histidine_biosynthesis, coenzyme_A_biosynthesis, formaldehyde_oxidation were decreased at week 2 (Supplementary Figure S4E). Similarly, most of these changes seemed to be transient, while some others were lasting until week 8 (Supplementary Figure S4E-H).
To further compare the differential functional pathways between the two groups, we conducted LEfSe analysis to identify the discriminating functional pathways. Interestingly, the distinct pathways were only identi ed between the two groups at week 4. We predicted 9 different microbiota-associated functional pathways between group A and group B. Four pathways with increased abundance in group B were predicted, including superpathway of thiamin diphosphate biosynthesis, sulfate_reduction, coenzyme_A_biosynthesis, and N10 formyl tetrahydrofolate biosynthesis. In contrast, the abundances of 5 pathways, which were associated with superpathway_of_lipopolysaccharide_biosynthesis, polymyxin_resistance, starch degradation, CDP_diacylglycerol_biosynthesis, and mannan_degradation were signi cantly decreased compared with group A ( Figure 5A).
Furthermore, we investigated whether the differential microbial taxa between group A and group B were related to the predicted pathways to depict how speci c taxa was involved in the functional pathways. The Enterococcus was signi cantly enriched in group B after probiotics supplementation, and we found Enterococcus was related to ve of the nine predicted pathways at week 4. Speci cally, Enterococcus was positively associated with superpathway of thiamin diphosphate biosynthesis, N10 formyl tetrahydrofolate biosynthesis, and coenzyme_A_biosynthesis, while negatively related to superpathway of lipopolysaccharide biosynthesis and polymyxin_resistance ( Figure 5B-F). In terms of the Collinsella, which was signi cantly decreased in group B, was positively related to starch degradation and CDP_diacylglycerol_biosynthesis, and mannan_degradation ( Figure 5G-I). Taken together, we hypothesized that the altered gut microbiota after eradication and probiotics supplementation might contribute to the various metabolic or clinical features through these microbiota related pathways.

Gut microbiome correlated with patients' features
Since various factors such as diet, age, alcohol, smoke, BMI might have impact on gut microbiota, we sought to investigate the relationship of gut microbiome and these factors after eradication. We incorporated the following variables into analysis: age, gender, BMI, alcohol, smoking, exercise. We found that age, alcohol and BMI were closely correlated with the microbial community diversity. We further identi ed speci c bacterial taxa in group A and group B by dividing patients into different subgroups based on these factors. The analysis showed that Streptococcus, Dorea, Selenomonas, and Mogibacterium were signi cantly increased, while Paraprevotella and Turicibacter decreased in elder patients in group A ( Figure 6A). However, Ruminococcus was signi cantly increased with a decrease of Megasphaera in elder patients in group B ( Figure 6A). Compared to non-drinkers, Gemmiger and Collinsella were increase after eradication, while Sporosarcina, Alloscardovia, and Moryella were enriched after probiotics supplementation ( Figure 6B). In terms of BMI, we found that Prevotella, Bilophila, Dialister, and Eikenella were markedly enriched after eradication, while Megasphaera, Blautia, and Faecalibacterium were increased after probiotics supplementation in obesity patients ( Figure 6C). These results indicated that age, BMI and alcohol had an impact on gut microbiota after H. pylori eradication, and probiotics supplementation might partially eliminate the adverse effect in patients with these speci c features.

Discussion
In this study, we performed a multicenter, randomized trial to show the distinct effects of bismuth quadruple therapy and probiotics (Medilac-S) supplementation on gut microbiota. Various studies have shown that the administration of antibiotics reduces the diversity of the gut microbiota [14][15][16]. Alpha diversity decreased 1 week post-eradication therapy and was restored to almost pre-eradication levels 8 weeks later [22]. Consistent with these studies, our data also presented a signi cant disruption of gut microbiota at the end (week 2) of eradication therapy. We then observed a trend of gradual restoration with time from week 4 to week 8, and the alpha and beta diversities were almost restored at week 8 in both groups. Unlike the previous study that probiotics could maintain the diversity of gut microbiota after eradication therapy [14], we found that probiotics supplementation failed to increase or maintain the diversity after H. pylori eradication.
In accordance with previous reports [22,23], in our current study, the gut microbiota before H. pylori eradication therapy predominantly contained commensal microbes of the phyla Firmicutes, Bacteroidetes, and Proteobacteria. Different eradication regimens containing different antibiotics might exert distinct effects on gut microbiota [15]. Oh, et al showed that the relative abundance of Firmicutes decreased and that of Proteobacteria increased immediately after triple therapy [14]. Hsu and colleagues showed that the relative abundance of Proteobacteria increased, whereas that of Bacteroidetes, Actinobacteria, and Verrucomicrobia decreased immediately after bismuth quadruple therapy containing pantoprazole, bismuth tripotassium dicitrate, tetracycline, and metronidazole [8]. Another study showed that reverse therapy containing pantoprazole, amoxicillin, clarithromycin, and metronidazole reduced the relative abundances of Firmicutes and Actinobacteria, while increased the abundance of Proteobacteria [24]. The enrolled patients in our present study received a 14-day bismuth quadruple therapy consisting of esomeprazole, amoxicillin, furazolidone, and bismuth potassium citrate. We observed that the relative abundance of Proteobacteria was signi cantly increased, while the abundance of Firmicutes, Bacterioidetes, Verrucomicrobia, and Actinobacteria was signi cantly decreased 2 weeks after treatment, which almost returned to the baseline levels at week 8. Amoxicillin and clarithromycin was supposed to contribute to the reduction of Firmicutes and Actinobacteria following eradication [24]. Our data revealed that Bacterioidetes and Verrucomicrobia were also signi cantly reduced, which might be attributed to furazolidone. The dramatic increase in the relative abundance of Proteobacteria, a major phylum of gram-negative bacteria including Escherichia, Proteus, Salmonella, Klebsiella, and Morganella, was observed after bismuth quadruple therapy. These discrepant observations may be explained partially by different eradication regimens, drug doses and treatment duration lengths. In addition, other factors, such as dietary habit, previous antibiotic treatment history, and individual differences in the absorption rate for antibiotics, can affect the in uence of eradication therapy on gut microbiota [25]. In our present study, we evaluated the food intake and baseline characteristics in each subjects and observed there was no obvious difference in the diet components and baseline characteristics. In terms of phylum level, probiotics supplementation did not obviously change the microbiome composition after eradication therapy.
There were important taxonomic changes at the genus level after treatment. BQT treatment at week 2 was associated with decreased abundance of Bacteroides, Faecalibacterium, Roseburia, Lachnospira, Phascolarctobacterium, Bi dobacterium, and Butyricimonas, most of which are known to have bene cial effects, such as producing the short chain fatty acid butyrate. On the contrary, there was an increase in relative abundances of some detrimental bacteria, such as Klebsiella, Streptococcus, Fusobacterium, Prevotella, and Morganella. Since amoxicillin and furazolidone have limited activity against these bacteria, it is likely that these detrimental bacteria may rapidly increase due to inhibition of other commensal bacteria. At week 2 after eradication, probiotics supplementation did not markedly alter the microbiota composition, suggesting the impact of antibiotics surpassed the protective effect of probiotics. With the prolonged eradication time, most of these bene cial and detrimental bacteria alterations were restored by week 6 and week 8, except for the persistent increase of Klebsiella in BQT group, implying that BQT therapy could lead to a persistent antibiotics resistance of Klebsiella. However, probiotics supplementation rapidly decreased the enrichment of Klebsilla, suggesting that the concomitant use of probiotics might be bene cial to reduce conditioned bacteria and antibiotics resistance. At week 2 and week 4, we observed a colonization of Enterococcus and Bacillus, the main components of the probiotics Medilac-S, while disappeared at week 6 and week 8, suggesting the colonization of probiotics closely depends on the supplementation duration. Most importantly, probiotics supplementation increased the bene cial bacteria such as Oscillospira and Lactobacillales at week 2 and week 4, which is reported to produce the short chain fatty acid butyrate, regulate host immune response and improve gastrointestinal symptoms [25,27]. Furthermore, the probiotics supplementation reduced the abundance of Dialister, Sutterella, and Collinsella, mainly contributing to the digestive disorder, in ammation, abnormal lipids metabolism and various metabolic syndrome [28][29][30][31]. However, we did not observed much valuable difference at week 6 and week 8 between the two groups, probably because the probiotics supplementation was abolished in this period.
The differences of functional pro les of gut microbiota after treatments were compared in antibiotic group and probiotic group. The proportion of pathways involved in starch degradation, glycolysis, and amino acid biosynthesis was decreased after treatment in both groups, whereas the proportion of pathways involved in the fatty acid oxidation and sucrose degradation increased. Various bacterial taxa such as Ruminococcus bromii and Bi dobacterium adolescentis are able to degrade starch [32], leading to an increase in speci c fermentation end products, in particular butyrate, promoting epithelial integrity and immune homeostasis [33]. During the metabolism of glycolysis and amino acid biosynthesis, these compounds act as donors for sugar residues in glycosylation reactions that produce polysaccharides, which are important constituents of the cell wall [34,35]. Moreover, the increased incidence of sucrose metabolism and fatty acid oxidation after eradication therapy were closely associated with the obesity and metabolic syndrome [36,37], which was consistent with some previous data indicating H. pylori eradication contributed to the metabolic parameters changes [38,39]. These intriguing ndings suggest that H. pylori eradication therapy might bring some potential detrimental effect through gut microbiota alterations, whether these changes are associated with signi cant clinical outcomes should be assessed in future studies. Under this circumstance, we further compared the changes of functional pathways between antibiotics group and probiotics group. Surprisingly, we only observed the distinct differences between the two groups at week 4. The proportion of pathways involved in the lipopolysaccharide biosynthesis, polymyxin resistance was increased only in the antibiotics group, while the metabolic pathways associated with metabolism of cofactors and vitamins were enriched in probiotics group.
Enterococcus, the main components of probiotics, was positively correlated with thiamin diphosphate biosynthesis, tetrahydrofolate biosynthesis, and coenzyme A biosynthesis, while negative related to lipopolysaccharide biosynthesis, polymyxin resistance, suggesting probiotics supplementation might help to construct a bene cial pro le of gut microbiota after eradication therapy.
We found that the variations of gut microbiome patterns are associated with several factors such as age, alcohol and BMI after eradication therapy. Several taxa can be identi ed to differentiate these clinical characteristics. In terms of age, we observed that Streptococcus and Dorea were signi cantly enriched in older patients after eradication therapy, while probiotics supplementation abolished this enrichment. Streptococcus is important conditioned pathogen causing in ammation and sepsis, and Dorea is the important acrobacter aerogenes contributing to the incidence of irritable bowel syndrome [40,41]. For alcohol drinkers, Gemmiger and Collinsella were highly increased after eradication, which were associated with the abnormal lipid metabolism, obesity and metabolic syndrome [42,43]. Moreover, for obese patients, Prevotella, Bilophila, Dialister, which have been reported to be enriched in obsess individuals [44], were also increased after eradication therapy. Intriguingly, the probiotics supplementation could effectively restore the adverse microbiome abundance in patients with the above characteristics, implying that H. pylori eradication therapy might have the most impact on the aged, drinking and obese individuals, and probiotics supplementation would gain the maximum bene t for these subjects.
Nevertheless, there were some limitations to this study. First, although the strength of this study includes a large-scale, multicenter, randomized trial, longer duration of follow-up is needed to investigate the longterm effect of eradication therapy and probiotics in gut microbiota. Second, the changes in the species level and detailed function pro les after eradication therapy could not be assessed through 16S rRNA sequencing. Further whole genome shotgun sequencing would be needed.

Conclusions
Bismuth quadruple therapy leads to transient perturbation and changes in the gut microbiota immediately after therapy, most of these changes returned to pretreatment levels by 8 weeks. Although probiotics supplementation did not obviously change the microbiota diversity compared to the antibiotics group, probiotics supplementation during H. pylori eradication might be bene cial to reduce conditioned bacteria and antibiotics resistance, which is closely dependent on the supplementation duration. Bismuth quadruple therapy might bring some potential detrimental effect through gut microbiota alterations, and probiotics supplementation changed the functional potential of gut microbiome, which is supposed to be bene cial to human health. Collectively, our results proposed that probiotics are bene cial for patients during H. pylori eradication, especially for patients with older age, alcohol drinking, and obesity, which might obtain the maximum bene ts.

Figure 2
Comparison of the diversity between the two groups after eradication therapy. Alpha diversity (A-C) indices alterations were compared between Group A and Group B before treatment, at week 2, at week 4, at week 6, and at week 8. Beta diversity (D-F; principal coordinate analysis) was compared between the two groups at different time periods.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.