基因芯片诊断耐多药结核病的临床多中心研究

A multicenter evaluation of a biochip system for detection of rifampin and isoniozid resistance in clinic strains of Mycobacterium tuberculosis

目的评估基因芯片法检测耐多药结核临床分离株的临床意义。方法采取分层抽样的方法分别从北京胸科医院、同济大学附属上海市肺科医院和广州市胸科医院保存的临床菌株库的耐药组和敏感组中，随机抽取利福平耐药株800株，异烟肼耐药株797株，耐多药株791株，利福平／异烟肼双敏感380株。用基因芯片法检测包括rpoB基因的511（T→C）、513（A→C，C→A）、516（G→T，A→T，A→G）、526（C→T，C→G，A→T，A→G）、531（C→T，C→G）、533（T→C）位点、katG的315（G→C，G→A）位点和inhA的-15（C→T）位点的耐药突变。以绝对浓度法药敏结果为金标准，计算基因芯片法的符合率、敏感度和特异度。同时对基因芯片法的核酸扩增产物进行测序，以验证基因芯片对核酸序列检测的准确性。结果以绝对浓度法药敏结果作为标准，基因芯片法检测利福平、异烟肼耐药和耐多药的符合率分别是93．7％（1108／1183）、83．8％（994／1186）、82．4％（975／1183）。检测利福平耐药的敏感度为92．0％（733／797），特异度为97．2％（375／386）；检测异烟肼的敏感度为77．4％（617／797），特异度为96．9％（377／389）；检测耐多药的敏感度为74．6％（588／788），特异度为98．0％（387／395）。在利福平基因芯片检测为突变的菌株中，突变频率最高的位点是531（TCG），突变率为64．5％（480／744）；在katG／inhA突变菌株中，基因芯片检测为katG315（AGC）单突变的为77．4％（487／629）；且与测序结果基本一致，仅有5例菌株中不完全相符。其中1株异烟肼耐药菌基因芯片法检测为katG315（G→C）突变，而测序结果为野生型，其余4株为基因芯片法未包含的突变类型。结论基因芯片法可快速可靠地检测结核临床分离株利福平和异烟肼的耐药性，有望在临床诊断中广泛应用。

Objective To evaluate a rapid biochip system for the determination of multidrug- resistant tuberculosis （MDR-TB） in Mycobacterium tuberculosis isolates. Methods A total of 1 186 clinical strains, including 800 rifampin （RFP） resistant isolates, 797 isoniozid （INH）resistant isolates, 791 MDR- TB and 380 susceptible strains, were selected from Beijing Chest Hospital, Shanghai Pulmonary Hospital and Guangzhou Chest Hospital respectively using stratified sampling method. Biochips were used to detect loci of rpoB 511 （T→C）, 513 （A→C, C→A）, 516 （G→T, A→T, A→G） , 526 （C→T, C→G, A→T, A→G）, 531 （C→T, C→G）, 533 （ W→C）, katG 315 （ G→C, G→A） and inhA -15 （ C→T）. Absolute concentration drug susceptibility test of RFP and INH were performed to serve as the gold standard to calculate susceptibility, specificity and overall concordance of biochip test. All polymerase chain reaction （PCR） products were sequenced to confirm the mutations. Results The concordances between the biochip system and absolute concentration drug susceptibility test were 93.7% （ 1 108/1 183） for RFP, 83.8% （994/1 186） for INH and 82.4% （975/1 183） for MDR-TB. Compared with absolute concentration drug susceptibility test, the biochip method displayed a sensitivity of 92. 0% （733/797） and 77.4% （617/797） and a specificity of 97. 2% （375/386） and 96. 9% （377/389） for RIF and INH, respectively. For MDR- TB, the biochip system reached a sensitivity of 74. 6% （588/788） and a specificity of 98.0% （ 387/395 ）. Among rpoB mutants, mutations were mostly detected at codon 531 [64. 5% （480/744）1. In stains with mutations in katG or inhA, 77.4% （487/629） had mutation at codon 315 （TCG） of katG only. The sequencing results had a high concordance with that of the biochip method. There were slight differences in 5 strains, among which one strain was detected by biochip as katG 315 （G→C） mutant, but was identified by sequencing as wild type, and mutation types other than those detected by the biochip were confirmed in the other 4 strains by sequencing. Conehasion This biochip system is adapted for extensive application in clinical diagnosis, as it allows fast and reliable detection of resistance to isoniazid and rifampin in tuberculosis clinical isolates.