全基因组耐药组分析预测铜绿假单胞菌药敏表型
本研究基于全基因组序列耐药组分析,开发并验证铜绿假单胞菌基因型耐药评分,用于预测抗菌药物敏感性表型。对100株野生型菌株测序定义40个染色体耐药基因的自然多态性目录,并在204株随机选取的临床分离株中测试评分预测表型的能力。基因型评分预测敏感性的能力很高(95-100%),但预测耐药性的能力因药物而异,美罗培南临床耐药预测率仅46.1%,而头孢他啶/阿维巴坦耐药预测率达100%。
要点
- 基因型耐药评分预测敏感性的能力高(95-100%),但预测耐药性的能力因药物差异大。
- 美罗培南临床耐药预测率低(46.1%),头孢他啶/阿维巴坦耐药预测率达100%。
- 野生型菌株中亦发现耐药突变和高风险克隆(如ST235),需警惕。
核心结论
- 基因型耐药评分与药敏表型总体相关性良好,但预测耐药性的能力因药物而异,需进一步优化评分系统。
研究设计与样本
- 基于40个染色体基因突变驱动耐药和水平获得性耐药(Resfinder)分析,开发针对头孢他啶、头孢他啶/阿维巴坦、美罗培南、环丙沙星和妥布霉素的基因型耐药评分系统。
- 对2017年多中心研究中来自51家医院的100株野生型菌株进行全基因组测序,定义40个染色体耐药基因的自然多态性目录。
- 在从51家医院随机选取的204株分离株(每家医院4株)中测试基因型评分预测药敏表型的能力。
主要发现
- 100株野生型菌株分析得到40个基因中455个自然多态性。
- 少数野生型菌株中亦发现耐药突变和高风险克隆(如ST235)。
- 基因型评分(<0.5)预测表型敏感(美罗培南为S+I)的能力很高(95-100%)。
- 基因型评分(≥1)预测耐药的能力因药物而异:美罗培南临床耐药预测率46.1%(18/39),头孢他啶/阿维巴坦耐药预测率100%(7/7)。
关键数值与推荐
- 基因型耐药评分范围0(无效应)至1(EUCAST临床耐药)。
- 预测敏感性阈值:评分<0.5;预测耐药阈值:评分≥1。
- 美罗培南敏感性定义为S+I。
临床怎么用
- 基因型耐药评分可用于预测铜绿假单胞菌对部分抗菌药物的敏感性,但预测耐药性需谨慎,尤其美罗培南。
- 临床应用中需结合表型药敏试验验证,尤其对于耐药预测率低的药物。
局限性
- 本研究为概念验证研究,评分系统仍有改进空间。
- 需进一步细化评分系统、实现自动化并在大型国际队列中验证。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVES: The aim was to develop and validate a Pseudomonas aeruginosa genotypic resistance score, based on analysis of the whole genome sequence resistome, to predict antimicrobial susceptibility phenotypes. METHODS: A scoring system based on the analysis of mutation-driven resistance in 40 chromosomal genes and horizontally acquired resistance (Resfinder) was developed for ceftazidime, ceftolozane/tazobactam, meropenem, ciprofloxacin and tobramycin. Resistance genes/mutations were scored from 0 (no effect) to 1 (EUCAST clinical resistance). One hundred wild-type strains obtained from 51 different hospitals during a 2017 multicentre study were fully sequenced and analysed in order to define a catalogue of natural polymorphisms in the 40 chromosomal resistance genes. The capacity of genotypic score to predict the susceptibility phenotype was tested in 204 isolates randomly selected from the 51 hospitals (four from each hospital). RESULTS: The analysis of the 100 wild-type isolates yielded a catalogue of 455 natural polymorphisms in the 40 genes involved in mutational resistance. However, resistance mutations and high-risk clones (such as ST235) were also documented among a few wild-type isolates. Overall, the capacity of the genotypic score (<0.5) for predicting phenotypic susceptibility (S + I in the case of meropenem) was very high (95-100%). In contrast, the capacity of the genotypic score to predict resistance (≥1) was far more variable depending on the agent. Prediction of meropenem clinical resistance was particularly low (18/39, 46.1%), whereas it classified clinical ceftolozane/tazobactam resistance in 100% (7/7) of cases. DISCUSSION: Although a margin for improvement was evidenced in this proof of concept study, an overall good correlation between the genotypic resistance score and the susceptibility profile was documented. Further refining of the scoring system, automatization and testing of large international cohorts should follow.
引用
- Cortes-Lara S, Barrio-Tofiño ED, López-Causapé C, et al. Predicting Pseudomonas aeruginosa susceptibility phenotypes from whole genome sequence resistome analysis. Clin Microbiol Infect. 2021. PMID: 34015532. doi:10.1016/j.cmi.2021.05.011
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。