Reliability of Mycobacterial Growth Indicator Tube (MGIT) 960 for the detection of isoniazid resistance in a tuberculosis endemic setting

Sir,

Isoniazid (INH) is one of the forerunners in the treatment of tuberculosis (TB). Resistance to INH along with that of rifampicin (RIF) is considered multidrug resistance (MDR)1. Several methods have been validated for detection of MDR-TB. One of the promising methods well evaluated and accepted in varying settings is the fluorimetry based liquid culture detection system, Mycobacterial Growth Indicator Tube (MGIT 960) (Becton and Dickinson, USA)234. However, reports of “false resistance” to INH at a concentration of 0.1 μg/ml recommended by the manufacturer have been documented567. Studies suggest that use of a higher concentration (0.4 μg/ml) provides better distinction of the “false resistant” (FR) isolates589. Being a vital drug in the treatment of tuberculosis, such discrepancies might render treatment ineffective. In the current study, we assessed the reliability of the lone use of the manufacturer recommended drug concentration of INH (0.1 μg/ml) with respect to conventional minimum inhibitory concentration (MIC) method on Lowenstein- Jensen (LJ) medium.

The study conducted in the department of Bacteriology, National Institute for Research in Tuberculosis, Chennai, India, included two batches of cultures. The first batch consisted of 101 Mycobacterium tuberculosis isolates obtained from new (n=30) and 71 previously treated pulmonary TB patients (inclusive of 45 Category I and 26 Category IV failure). These isolates were tested for drug susceptibility (DST) to INH at 0.1μg/ml concentration by MGIT 960 system. Conventional DST by MIC method on L-J medium at concentrations 0.2, 1.0, and 5.0 μg/ml was performed for all the isolates10.

The second batch had 60 M. tuberculosis isolates received from the Institute of Tropical Medicine, Belgium, as a part of routine external quality assurance (EQA). The isolates were tested for susceptibility to INH by MIC method and proportion susceptibility test (PST) at a concentration of 0.2 μg/ml10, BACTEC 460 and MGIT 960 at a concentration of 0.1 μg/ml. DST by liquid culture systems was performed according to the manufacturer's protocol. Appropriate numbers of duplicates were included in MGIT 960 and BACTEC 460 as internal quality controls. Statistical methods using Chi square testing to assess the performance parameters of MGIT 960 were performed using SPSS software version 14.0, USA.

The results indicated high sensitivity (>91%), specificity (>95%) and accuracy (>92%) for detection of INH resistance by MGIT 960 in comparison with conventional MIC method (Table I) for the 101 isolates tested. Four isolates exhibited discordant results by MGIT 960 of which two were categorized as false resistant (FR) and two as false susceptible (FS).

Table I Comparison of INH susceptibility between MGIT 960 and conventional MIC method among clinical isolates (n=101)

Comparison of MGIT 960 with phenotypic methods (MIC and PST) and BACTEC 460 for EQA isolates showed a good concordance in sensitivity and specificity (Table II). MGIT 960 indicated three FR and FS isolates in comparison with MIC and PST. One of the three FS isolates in MGIT 960 showed intermediate resistance (IR) phenotype by conventional MIC method. According to Van Deun11 reporting IR strain as susceptible is acceptable in case of INH. Similar case of FS isolate was observed by Abe et al12 with phenotypic resistance at 1.0 μg/ml. Two of the FS isolates showed MIC of 5 μg/ml by conventional MIC method. MGIT 960 in comparison with BACTEC 460 demonstrated three FS isolates and a single FR isolate. Results of the latter were resolved in accordance with MGIT 960 when compared with MIC and PST methods. The susceptibility pattern of duplicates was concordant with that of original.

Table II Comparison of INH susceptibility by MGIT 960, BACTEC 460 and phenotypic methods among external quality assurance (EQA) isolates (n = 60*)

Use of MGIT 960 in routine mycobacteriology has dramatically reduced the turn around time for detection of resistance thus paving way for early and accurate intervention13. Sensitivity and specificity of MGIT 960 for detection of INH resistance observed in our study was in accordance with earlier reports679. Studies indicate that MGIT 960 has a tendency to indicate more INH resistance rates than BACTEC 460 with increased chance for FR4814. False resistant and FS isolates were observed to a limited extent in the present study. The uneven distribution of the heterogeneous population in subculture might have resulted in a false susceptible result2. In addition, varying growth indices in MGIT 960 while preparing inocula could also contribute to such discrepancies at higher frequencies. This warrants further evaluation with a larger number of isolates with intermediate resistances and different ranges of growth indices.

Presence of micro clumps in the inoculum carrying uneven distribution of mycobacteria, seeding the inoculum using pipette which allows large clumps and difference in the DST procedure when performed at different time points may contribute to false resistance in MGIT 96014. Despite the discrepancies observed, the accuracy of MGIT 960 was acceptable (≥92%). Agreement between the methods was found to be high (>0.8).

One limitation in this study was that the status of discrepant isolates was not reconfirmed by INH at 0.4μg/ml concentration. Validation of FR using a higher drug concentration was thought to be superfluous with high accuracy of MGIT 960 and limited discrepancy between phenotypic methods. In an unrelated study by the authors15, existence of intermediate resistance (IR) to INH (data not shown) that could lead to discrepancy between MGIT 960 and conventional methods was found to be minimal in our setting. High level isoniazid resistance (≥5μg/ml) due to mutation in katG gene was observed within the subcontinent15. Hence, with less number of IR to INH, DST by MGIT 960 can be performed at manufacturer's recommended concentration of 0.1 μg/ml in the present clinical setting.

Currently, MGIT 960 being considered the gold standard in liquid culture system and introduced for diagnosis and DST of M. tuberculosis under Revised National TB Control Programme16, laboratory specific assessment of the recommended drug concentrations at regular time intervals is required to eliminate inconsistency and improve reliability.