Objective To evaluate the differences among microarray chip method, BACTEC MGIT 960 method (MGIT960) and Löwenstein-Jensen (L-J) proportion method in the detection ofisoniazid (INH) and rifampicin (RFP) drug sensitivity test of Mycobacterium tuberculosis (MTB) forguiding clinical medication.
Methods The confirmed and suspected samples of TB inpatients and outpatients in the clinical departments of our hospital from March 2016 to November 2017 were chosen and detected by microarray, MGIT960 and L-J. From which 446 samples of patients with positive results were selected for INH and RFP drug sensitivity test, and the differences of the results of the three methods were compared.
Results The results of MGIT960 and L-J were used as references respectively, the sensitivity of microarray chip to detect INH resistance of MTB were 84.8% and 83.5%, and the specificity were 99.7% and 98.5%, Kappa values were 0.885 and 0.855 respectively, while the sensitivity of microarray chip to detect RFP resistance of MTB were 96.2% and 93.7%, and the specificity were 93.8% and 93.5%, Kappa values were 0.817 and 0.796 respectively. The results showed that the microarray chip method and the other two methods had excellent consistency in the detection of INH and RFP resistance. In addition, MTB samples with inconsistent drug susceptibility of MGIT960 and L-J proportion method were detected by microarray chip method. The results showed that the coincidence rates of RFP resistance were 77.8% and 88.9%, while INH were 100.0% and 75.0% respectively. These results suggested that there may be heterogeneous drug resistance in MTB samples with inconsistent drug susceptibility between MGIT960 and L-J method.
Conclusion Compared with MGIT960 and L-J proportion method, microarray chip method has excellent consistency in the detection of INH and RFP resistance, especially, it has the advantages of accuracy, rapidity, high sensitivity and specificity. Furthermore, microarray chip method can be used to detect clinical samples with inconsistent phenotype susceptibility, which could be verified and provided better laboratory basis for treatment of the clinical heterogeneous drug resistance.
摘要： 目的 评价微阵列芯片法与BACTEC MGIT 960液体药敏 (简称 MGIT960 药敏) 、罗氏比例法药敏 (简称罗氏 比例法) 在结核分枝杆菌 (MTB) 对异烟肼 (INH) 、利福平 (RFP) 药敏试验检测方面的差异, 用于指导临床用药。 方法 选择医院临床科室2016年3月—2017年11月期间确诊及疑似结核病的住院和门诊患者标本, 同时采用微阵列 芯片法、 MGIT960 药敏和罗氏比例法对其进行结核分枝杆菌检测, 选择3种方法检测均为阳性的 446 例患者标本进行 异烟肼和利福平药敏检测, 比较三者检测结果的差异性。 结果 分别以 MGIT960 药敏结果和罗氏比例法为参照, 微阵 列芯片法检测 MTB 对异烟肼耐药性的灵敏度分别为84.8%和83.5%、特异度分别为99.7%和98.5%、 Kappa 值分别0.885 和0.855, 对利福平耐药性的灵敏度分别为96.2%和93.7%、特异度分别为93.8%和93.5%、 Kappa 值分别为0.817 和 0.796, 表明微阵列芯片法和另外两种方法对异烟肼和利福平耐药性的检测具有极好的一致性; 此外, MGIT960 药敏和 罗氏比例法药敏不一致的MTB样本采用微阵列芯片法检测, 结果显示, MGIT960 药敏和罗氏比例法分别与微阵列芯片 法比较, 利福平的耐药性符合率分别为77.8%和88.9%, 异烟肼的耐药性符合率分别为100.0%和75.0%, 表明两种表型 药敏不一致的MTB样本可能存在异质性耐药。 结论 微阵列芯片法与 MGIT960 药敏和罗氏比例法药敏比较, 对异烟 肼和利福平耐药性检测具有极好的一致性, 具有快速、精准、较高的灵敏度和特异度, 对于表型药敏不一致的临床样本 采用微阵列芯片法检测, 可对其进行验证并为临床上异质性耐药的治疗提供更好的实验室依据。