Objective To investigate the characteristic changes in gut microbiota before treatment in children with Helicobacter pylori (H. pylori) infection, and to analyze the predictive value of differential gut microbiota for treatment outcomes and drug resistance.

Methods A total of 168 children with H. pylori infection were enrolled and treated with standard quadruple therapy for 14 days. Based on eradication efficacy, they were divided into eradication success group (118 cases) and eradication failure group (50 cases); based on drug susceptibility results, they were divided into sensitive group (112 cases) and resistant group (56 cases). Fecal samples were collected from the children before treatment, and gut microbiota diversity and composition were analyzed using 16S rRNA high-throughput sequencing. LEfSe was employed to screen for differential bacterial genera. The receiver operating characteristic (ROC) curve was used to evaluate the performance of differential gut microbiota in predicting treatment outcomes and drug resistance in children with H. pylori infection.

Results The Shannon index and Simpson index of gut microbiota before treatment in the eradication failure group were lower than those in the eradication success group, with statistically significant differences (P < 0.001). The Shannon index of gut microbiota before treatment in the resistant group was lower than that in the sensitive group, with statistically significant difference (P < 0.001). Significant differences in gut microbiota structure were observed between the eradication success and eradication failure groups, as well as between the sensitive and resistant groups (P < 0.01). The abundances of Bifidobacterium and Lactobacillus in the eradication success group were higher than those in the eradication failure group, while the abundances of Bifidobacterium and Akkermansia in the sensitive group were higher than those in the resistant group (LDA>3.0, P < 0.05). The abundance of opportunistic pathogenic bacteria such as Enterococcus in the eradication failure group and the drug-resistant group increased (LDA>3.0, P < 0.05). The area under the curve (AUC) for predicting eradication success using Bifidobacterium combined with Lactobacillus was 0.885 (95%CI, 0.832 to 0.938); the AUC forpredicting H. pylori sensitivity using Bifidobacterium combined with Akkermansia was 0.912 (95%CI, 0.865 to0.959). The net benefit of the combined models was superior to that of single bacterial genus prediction.

Conclusion Characteristic changes exist in the gut microbiota of children with H. pylori infection before treatment, among which Bifidobacterium, Lactobacillus and Akkermansia demonstrate good predictive value for treatment outcomes and drug resistance.