【作者】 刘敏；

【导师】 陈光侠；

【作者基本信息】 中国矿业大学 ， 生物与医药（专业学位）， 2024， 硕士

【摘要】 背景幽门螺杆菌（Helicobacter Pylori,HP）在世界范围内具有较高的感染率,其感染易引发许多消化系统疾病与消化系统外疾病,严重威胁人类健康,HP的预防和治疗十分重要,然而HP相关抗生素的耐药率呈现增加的趋势。目的为了预防HP感染、探究疗效高的HP方案以减少耐药性的出现和提高医疗效率:（1）建立成人感染HP的临床预测模型并筛选感染HP的危险因素。（2）探索社区人群中HP耐药类型及耐药位点的分布以指导制定HP根除方案。（3）探究新型伏诺拉生阿莫西林二联方案和传统雷贝拉唑四联方案的疗效与安全性。（4）分析二联方案和四联方案根除前后肠道菌群差异。方法（1）随机纳入2022年7月至2023年7月徐州第一人民医院就诊的704例患者为研究对象,收集患者的30个特征,采用特征筛选、平衡数据、划分数据集、模型训练与调参等方法构建基于机器学习的成人HP感染预测模型并分析危险因素。（2）随机纳入2023年5月至2023年7月徐州市铜山区社区的198例居民为研究对象,收集粪便样本,利用荧光PCR探针法检测HP;多重PCR和核酸质谱法检测HP耐药基因,掌握社区人群耐药情况并指导制定HP根除方案。（3）随机纳入2023年3月至2023年9月于徐州第一人民医院就诊的171例HP阳性患者,随机分为A组（二联组,伏诺拉生20 mg一日两次、阿莫西林1 g一日三次,用药十四天）和B组（四联组,雷贝拉唑20 mg一日两次、胶体果胶铋300 mg一日两次、呋喃唑酮100 mg一日两次、阿莫西林1 g一日两次,用药十四天）,治疗结束至少4周后复测¹³C呼气试验,分析两组治疗方案的根除率与不良反应差异。（4）随机纳入2023年3月至2023年9月于徐州第一人民医院就诊的30例二联治疗患者（A组）、28例四联治疗患者（B组）、20例阴性患者（C组）进行肠道菌群研究,经样本收集、DNA提取、PCR扩增、16S二代测序、生信分析等方法分析二联和四联方案根除HP前后肠道菌群的差异。结果（1）共纳入677名医院患者,该人群HP感染率为62.33%。提出的XGboost模型性能优于逻辑回归、随机森林等机器学习模型,准确度、召回率、F1分数、AUC值最高分别为78.08%、74.27%、77.27%、0.861,精确度为80.52%。家庭成员HP感染、非萎缩性胃炎、农村居住、饭前便后洗手习惯差是HP感染的危险因素;城区居住是HP感染的保护因素。（2）最终纳入180例社区居民的粪便样本,该人群HP感染率为39.44%,HP阳性组和阴性组在性别上有显著性差异（P<0.05）。社区HP基因型耐药率为76.06%,多重耐药率为12.68%;耐药率最高的抗生素为克拉霉素23S（66.20%）,克拉霉素23S耐药均在A2143G位点;其次是喹诺酮类gyr A（21.13%）,在N87K位点耐药占比量最高为66.67%。仅有1例呋喃唑酮pord耐药,未检测到阿莫西林PBP1A和四环素16S的耐药。（3）最终纳入78例A组患者和77例B组患者。意向性治疗分析中A组和B组的根除率无显著性差异（P>0.05,分别为84.6%和74.0%）,符合方案集分析中A组和B组的根除率差异显著（P<0.05,分别为97.1%和86.4%）。两组不良反应情况无显著性差异,均出现了恶心、腹胀、腹痛等不良反应且程度均较轻,A组还出现了腹泻的情况;大部分患者复查时不良反应消失。（4）A组21对样本、B组20对样本、C组20例样本纳入肠道菌群分析,HP阳性和阴性患者的肠道菌群在丰度排名前十的门、科、属水平、alpha多样性、beta多样性、预测功能方面存在显著性差异（P<0.05）,二联和四联方案治疗前后患者的肠道菌群在丰度排名前十的门、科、属水平、alpha多样性、beta多样性上无显著性差异（P>0.05）。四联组治疗前后肠道菌群存在差异显著的预测功能（P<0.05,假定蛋白酶、DNA（胞嘧啶-5）-甲基转移酶1等）,二联组治疗前后肠道菌群预测功能无显著性差异（P>0.05）。结论（1）基于XGboost提出的成人HP感染预测模型性能最高,可为鉴定幽门螺杆菌易感人群和预防幽门螺杆菌感染提供科学依据（2）呋喃唑酮、阿莫西林、四环素的HP耐药率低,与其相关根除方案可推荐使用（3）伏诺拉生-阿莫西林二联方案治疗效果优于四联方案（4）伏诺拉生-阿莫西林二联方案治疗对肠道菌群的影响小于四联方案。该论文有图20幅,表23个,参考文献173篇。更多还原

【Abstract】 Background Helicobacter pylori（HP）infection is highly prevalent worldwide and is associated with various digestive system and extra-digestive diseases,posing a significant threat to human health.The prevention and treatment of HP infection are crucial;however,the resistance to HP-related antibiotics is increasing.Objectives To prevent HP infection and explore highly effective HP regimens that reduce the emergence of drug resistance and improve medical efficiency,this study aims to:（1）Establish a clinical prediction model for adult HP infection and identify risk factors for HP infection.（2）Investigate the distribution of HP resistance types and drug resistance loci in the community population to guide the development of HP eradication program.（3）Evaluate the efficacy and safety of the novel vonoprazan and amoxicillin double regimen compared to the traditional rabeprazole sodium quadruple regimen.（4）Analyze the differences in gut microbiota before and after HP eradication.Methods（1）A total of 704 patients who were treated at the First People’s Hospital of Xuzhou from July 2022 to July 2023 were randomly included in the study,and 30 characteristics of patients were collected.Machine learning-based techniques such as feature screening,data balancing,data set division,model training,and parameter adjustment will be employed to construct a prediction model for adult HP infection and analyze risk factors.（2）A total of 198 residents from Tongshan District of Xuzhou from May 2023 to July 2023 were randomly included in the study.Stool samples were collected and HP was detected by fluorescent PCR probe method.Multiplex PCR and nucleic acid mass spectrometry were used to detect HP drug resistance genes to understand the drug resistance of community population and guide the development of HP eradication program.（3）171 HP positive patients admitted to the First People’s Hospital of Xuzhou from March 2023 to September 2023 were randomly divided into group A（double therapy group,vonoprazan 20mg twice a day and amoxicillin 1g three times a day for 14 days）and group B（quadruple therapy group,rabeprazole 20mg twice a day,colloidal bismuth pectin 300mg twice a day,furazolidone 100mg twice a day,amoxicillin 1g twice a day,treatment for 14 days）.At least 4 weeks after the end of treatment,¹³C breath test was repeated,and the eradication rate and adverse reactions were compared between the two groups.（4）A total of 30 patients with dual therapy,28 patients with quadruple therapy and 20HP-negative patients in the First People’s Hospital of Xuzhou from March 2023 to September 2023 were randomly enrolled for the study of intestinal flora.The differences in intestinal flora before and after HP eradication with dual therapy and quadruple therapy were analyzed by sample collection,DNA extraction,PCR amplification,16S next generation sequencing and bioinformatics analysis.Results（1）A total of 677 patients were enrolled,and 30 patients’characteristics were collected.The rate of HP infection was 62.33%.The performance of the proposed XGboost model was better than that of machine learning models such as Logistic Regression and Random Forest,with the highest accuracy,recall,F1 score,and AUC value of 78.08%,74.27%,77.27%,and 0.861,respectively,and the precision was 80.52%.Risk factors for HP infection included HP infection in family members,non-atrophic gastritis,living in rural areas,and poor habits of washing hands before meals and after toileting.Living in urban areas was a protective factor for HP infection.（2）A total of 180 stool samples were collected from community residents.The HP infection rate in this population was 39.44%,and there was notable difference in gender between HP positive and HP negative groups（P<0.05）.The genotypic resistance rate of HP in community was 76.06%,and the multi-drug resistance rate was 12.68%.The antibiotic with the highest resistance rate was clarithromycin 23S（66.20%）,and the resistance was found at the A2143G site.The second highest resistance rate was quinolone gyr A（21.13%）,with the N87K site showing the highest proportion of drug resistance（66.67%）.Only one case was resistant to furazolidone,and no resistance was detected for amoxicillin PBP1A and tetracycline 16S.（3）A total of 78 patients were enrolled in group A,while 77 patients were enrolled in group B.In the intention-to-treat analysis,there was no statistically significant difference observed in the eradication rates between the dual therapy and quadruple therapy groups（84.6%and 74.0%,respectively;P>0.05）.However,in the per-protocol analysis,a significant difference was found in the eradication rates between the dual therapy group and the quadruple therapy group（97.1%and 86.4%,respectively;P<0.05）.Both groups experienced mild adverse reactions such as nausea,abdominal distension,abdominal pain,but only diarrhea occurred in the dual therapy group.Importantly,there were no significant discrepancies noted in adverse reactions between the two groups during follow-up period.（4）Twenty-one pairs of samples from the HP positive group,20 pairs of samples from the HP negative group,and 20samples from the HP negative group were included in the analysis of gut microbiota.There were significant differences in the top ten phylum,family,genus level,alpha diversity,beta diversity,and predictive function between the HP positive and negative patients（P<0.05）.There were no significant differences in the abundance of the top ten phyla,families,genus,alpha diversity and beta diversity between the patients before and after dual therapy and quadruple therapy（P>0.05）.There was a significant difference in the predicted function of intestinal flora before and after treatment in the quadruple therapy group（P<0.05,DNA（cytosine-5）-methyltransferase 1,etc）,while there was no significant difference in the predicted function of intestinal flora before and after treatment in the double therapy group（P>0.05）.Conclusions（1）The prediction model of adult HP infection based on XGboost has the highest performance,which can provide a scientific basis for identifying the susceptible population of Helicobacter pylori and preventing Helicobacter pylori infection.（2）The prevalence of HP resistant to furazolidone,amoxicillin and tetracycline is low,and their related eradication regimens can be recommended.（3）The effect of vonoprazan-amoxicillin double therapy is better than that of quadruple therapy.（4）The effect of vonoprazan-amoxicillin double therapy on intestinal flora is less than that of quadruple therapy.There are 20 figures,23 tables and 173 references in this thesis.更多还原