基于Mantel检验的大肠埃希菌耐药率相关性可视化分析

Visual analysis of correlation between bacterial resistance rate of Escherichia coli based on Mantel test

关键词:  Mantel检验  抗菌药物  细菌耐药率  使用强度

Key words:  mantel test  antibacterial drugs  bacterial resistance rate  antibiotics use density

摘要:

目的 分析大肠埃希菌耐药性与抗菌药物使用强度的相关性，为临床抗菌药物监管提供依据。方法 从医院信息系统提取提取2017年1月-2023年6月的住院患者大肠埃希菌耐药率数据和22种抗菌药物的使用强度数据。采用Mantel检验对细菌耐药情况进行整体分析，并以R语言进行可视化呈现。结果 2017年1月-2023年6月大肠埃希菌对三代头孢菌素、喹诺酮类、三代头孢加酶制剂和碳青霉烯类耐药率均无明显增涨趋势，平均耐药率分别为53.0%、45.02%、6.75%和1.44%。对β内酰胺加酶制剂耐药率与头孢哌酮舒巴坦和哌拉西林他唑巴坦AUD相关性最强，同时与头孢美唑、头孢西丁、头孢噻肟、头孢呋辛、美洛西林和磺苄西林AUD显著相关；对三代头孢类、喹诺酮类和碳青霉烯类耐药率与其相应制剂的AUD相关性均不显著，但对三代头孢平均耐药率与阿莫西林克拉维酸钾和美洛西林AUD、对喹诺酮类耐药率与头孢唑林和青霉素AUD呈现了相关性。结论 细菌耐药率与抗菌药物用量之间存在的宏观量化关系较复杂，利用Mantel检验能够对AUD和BRR两组矩阵之间的相关性进行整体可视化分析，直观展现复杂多维数据的相关性，提高抗菌药物及细菌耐药监控效率，是制定抗感染治疗方案和抗菌药物管控方案有效手段。

Abstract:

OBJECTIVE To provide information for the rational use of antibacterials, the variation of Escherichia coli resistance as well as the relationship between the bacterial resistance rate(BRR) and antibiotics use density(AUD) of antibiotics were investigated. METHODS The original BRR of Escherichia coli and the AUD of 22 antibiotics from January 2017 to June 2023 were collected from the hospital information system. Mantel test was used to comprehensively analyze the correlation between bacterial drug resistance and visualized in R language. RESULTS The BRR of Escherichia coli to third-generation cephalosporins, quinolones, BLBLIs and carbapenems showed no obvious upward trend from January 2017 to June 2023, and the average were 53.0%, 45.02%, 6.75% and 1.44% respectively. The BRR of Escherichia coli to BLBLIs showed the strongest correlation with the AUD of cefoperazone/sulbactam and piperacillin/tazobactam, and was significantly correlated with the AUD of cefmetazole, cefoxitin, cefotaxime, cefuroxime, mezlocillin and sulbacillin. There was no significant correlation between AUD and drug resistance rate of the third generation cephalosporins, quinolones and carbapenems. However, the resistance rate to the third generation cephalosporins was correlated with the AUD of amoxicillin clavulanate potassium and mezlocillin, and the resistance rate to quinolones was correlated with the AUD of cefazolin and penicillin. CONCLUSION There is a complex macro quantitative relationship between the BRR and the AUD of antibiotics. The Mantel test can be used to visually analyze the correlation between the AUD and BRR matrices as a whole, intuitively demonstrating the correlation of complex multidimensional data and improving the efficiency of monitoring antibacterial drugs and bacterial resistance. It is an effective means to formulate anti infective treatment plans and antimicrobial drug control plans.