目的 了解广西壮族自治区结核分枝杆菌乙胺丁醇（EMB）耐药基因突变特征及其影响因素，为结核病的分子诊断和临床治疗用药提供依据。 方法 对2018-2019年广西30个结核病耐药监测点连续不间断收集的655株结核分枝杆菌（其中52株乙胺丁醇耐药菌株，603株乙胺丁醇敏感菌株）进行全基因组测序，分析乙胺丁醇耐药基因突变特征及其影响因素。 结果 655株结核分枝杆菌中，54株发生乙胺丁醇耐药基因突变，突变率为8.24%（54/655）。在比例法检测EMB耐药的52株菌株中，21株发生基因突变，突变率为40.38%（21/52）；603株EMB敏感株中，33株发生基因突变，突变率为5.47%（33/603），耐药株中的基因突变率高于敏感株（χ² = 77.133，P = 0.000）。EMB耐药表型与基因突变的符合率为40.38%（21/52），二者检测结果吻合度不高（Kappa值 = 0.343，P < 0.001）。54株发生基因突变的菌株，突变基因为embA、embB和embC，突变形式有20种，其中单位点突变29株，占53.70%（29/54），联合位点突变25株，占46.30%（25/54）。单位点突变中，突变比例最高的位点为embB306（35.19%），其次为embB497（5.56%）和embB406（3.70%）。联合位点突变中，突变比例最高的位点为embC270+embB378（22.22%），其次为embB306+embA-12（3.70%）。性别、抗结核治疗史、基因型以及是否耐多药可能与EMB基因突变有关（χ² =9.388，P = 0.004；χ² = 27.084，P = 0.000；χ² = 6.671，P = 0.010；χ² = 68.826，P = 0.000）。多因素logistic回归分析结果显示男性（OR = 6.150）、复治（OR = 2.636）和耐多药（OR = 7.333）可能是EMB耐药基因突变的危险因素，北京基因型可能是EMB耐药基因突变的保护因素（OR = 0.511）。 结论 结核分枝杆菌EMB耐药性与embA、embB和embC基因突变有关，EMB耐药表型与基因突变吻合率不高。对于男性、复治、耐多药以及非北京基因型菌株感染的结核病患者，应关注其EMB耐药基因突变的情况。

Objective To investigate the mutation characteristics and influencing factors of ethambutol （EMB） resistance gene of Mycobacterium tuberculosis in Guangxi Zhuang Autonomous region， and to provide evidence for molecular diagnosis and clinical treatment of tuberculosis. Methods A total of 655 strains of Mycobacterium tuberculosis （52 ethambutol resistant strains and 603 ethambutol sensitive strains） were collected continuously from 30 TB drug resistance monitoring sites in Guangxi in 2018-2019， and the mutation characteristics and influencing factors of ethambutol resistant genes were analyzed by whole genome sequencing. Results Among 655 strains of Mycobacterium tuberculosis， 54 strains had ethambutol drug resistance gene mutation， the mutation rate was 8.24% （54/655）. Among 52 EMB-resistant strains detected by proportional method， 21 had gene mutation， the mutation rate was 40.38% （21/52）， and 33 of 603 EMB-sensitive strains had gene mutation， the mutation rate was 5.47% （33/603）. The gene mutation rate in drug-resistant strains was higher than that in sensitive strains （χ ² =77.133， P = 0.000）. The coincidence rate of EMB drug resistance phenotype and gene mutation was 40.38%（21/52）， and the results of the two tests were not highly consistent （Kappa = 0.343， P < 0.001）. The mutant genes of 54 strains were embA， embB and embC， and there were 20 mutant forms， among which 29 were mutated at unit point， accounting for 53.70% （29/54）， and 25 were mutated at joint site， accounting for 46.30% （25/54）. Among the unit point mutations， embB306（35.19%） had the highest mutation proportion， followed by embB497 （5.56%） and embB406 （3.70%）. Among the joint site mutations， embC270+embB378 had the highest mutation proportion （22.22%），The second was embB306+embA-12 （3.70%）. Gender， anti-tuberculosis treatment history， genotype and MDR might be related to EMB gene mutation （χ ² = 9.388， P = 0.004； χ ² = 27.084， P = 0.000； χ ² = 6.671， P = 0.010； χ ² = 68.826， P = 0.000）. Multivariate logistic regression analysis showed that male （OR = 6.150）， retreatment （OR = 2.636） and multidrug resistance （OR = 7.333） may be risk factors for EMB resistance gene mutation， and Beijing genotype may be a protective factor for EMB resistance gene mutation （OR = 0.511）. Conclusion EMB resistance of Mycobacterium tuberculosis is related to embA， embB and embC gene mutations， and the incidence of EMB resistance phenotype is not high. For male， retreatable， MDR-resistant， and non-Beijing genotype TB patients， attention should be paid to the mutation of the EMB resistance gene.