Rapid susceptibility testing of Mycobacterium tuberculosis by the Mycobacteria Growth Indicator Tube (MGIT AST SIRE)
Corresponding author and reprint requests: M. Souleymane, Laboratoire de Bactériologie-Virologie, Faculté de Médecine et de Pharmacie, CHU A. Le Dantec de Dakar, BP 7325 Dakar-Sénégal
Tel: + 221 822 59 19/821 64 20 Fax: + 221 821 64 42
Objective To evaluate the reliability of the Mycobacteria Growth Indicator Tube (MGIT AST) for susceptibility testing of Mycobacterium tuberculosis.
Methods Seventy strains of M. tuberculosis were tested for susceptibility to streptomycin, isoniazid, rifampicin and ethambutol by comparing MGIT AST results to those obtained by the method of proportion (MOP) on Lowenstein–Jensen (LJ) and Middlebrook 7H10 media. The 7H10 MOP was considered the method of reference.
Results The turnaround time for MGIT AST was 6.2 days (5–10 days) and for MOP it was 18–21 days. With rifampicin, MGIT AST agreed for all isolates with both MOP. For streptomycin, MGIT AST and 7H10 MOP agreed for 64 isolates (91.4%); 61 were susceptible and three resistant. LJ MOP and 7H10 MOP agreed for 64 isolates (92.2%); 62 were susceptible and three resistant. With isoniazid, both MOP agreed for all isolates, while MGIT AST and 7H10 MOP had two discrepancies. For ethambutol, MGIT AST and 7H10 MOP were concordant for 66 isolates; 65 were susceptible and one resistant. Both MOP were concordant for 67 isolates; 66 were susceptible and one resistant.
Conclusions Based on these results, MGIT AST is a time-saving method and can be used as an alternative to the BACTEC System. MGIT AST is reliable as far as rifampicin and isoniazid are concerned; however, additional studies are needed for streptomycin and ethambutol.
Tuberculosis (TB) remains a major public health problem in both developing and developed countries. In Africa, despite national program facilities, the rate of TB continues to increase. The lack of accurate methods for diagnosis and the long delay in providing results are among the main reasons for this high rate.
Standard methods, methods of proportion on Middlebrook 7H10 agar and Lowenstein–Jensen (LJ), for determining the drug susceptibility of Mycobacterium tuberculosis require 3–4 weeks to complete, while patients with resistant organisms receive treatment with an ineffective drug regimen. This delay may lead to additional drug resistance and failure to control the disease.
Given the increasing rate of TB, there is a need to develop more rapid and efficient methods for mycobacterial tests (diagnosis, susceptibility testing, characterization, etc.). To face this challenge, experts from the Centers for Disease Control (CDC) proposed a turnaround time of 30 days from the receipt of a specimen in the laboratory to complete mycobacterial tests .
Of the commonly used methods, only the BACTEC 460 System can provide results within 30 days [2,3]. Unfortunately, both the cost and the management of radioactive waste limit the use of the BACTEC System.
Recently, Becton Dickinson has introduced the Mycobacteria Growth Indicator Tube (MGIT), a broth-based non-radiometric system, as an alternative to radiometric systems. MGIT is a 1 × 100 mm tube containing a modified Middlebrook 7H9 broth with an oxygen-quenching fluorescent indicator compound embedded in silicone at the bottom. When the actively growing mycobacteria consume the oxygen dissolved in the broth, the indicator fluoresces when exposed to 365-nm UV light from a transilluminator. By adding M. tuberculosis suspension to a drug-containing MGIT, an antimycobacterial susceptibility test (AST) can be performed by comparing the growth of the drug-containing MGIT with that of the growth control (GC) (drug-free MGIT). Many studies have reported the ability of MGIT to shorten the time for both diagnoses and the reporting of susceptibility testing [3–7].
The purpose of this study was to assess the reliability of MGIT AST as a rapid drug susceptibility testing method in a developing country where LJ is the commonly used medium.
MATERIALS and METHODS
Seventy-one strains of M. tuberculosis isolated from clinical specimens of individual patient in Aristide Le Dantec Hospital in Dakar (Senegal) and identified by conventional biochemical testing [8,9] were tested for drug susceptibility with the following first-line drugs: streptomycin, isoniazid, rifampicin and ethambutol. Susceptibility was performed in MGIT AST, and on LJ and Middlebrook 7H10 agar medium, by the 1% method of proportion (MOP) as described previously [8,10,11]. The MOP on Middlebrook 7H10 agar medium was used as reference method. M. tuberculosis H37Rv (ATCC 27294) and an in-house M. tuberculosis strain resistant to the four drugs were included each time the susceptibility testing was performed as quality control (QC) strains and to check for the reproducibility of the method.
For inoculum preparation, colonies were scraped from LJ slants and transferred to a 16 × 100 mm sterile glass tube containing 3 mL of Middlebrook 7H9 broth and eight glass beads. Tubes were vigorously agitated on a vortex mixer for 3–2 min, and clumps were allowed to settle for 20 min. The supernatant was removed and transferred to another sterile glass tube, and clumps were again allowed to settle for 15 min. The supernatant from this tube was removed and transferred to a final sterile glass tube and adjusted to match the densities of 1 and 0.5 standard McFarland suspension. MGIT AST inoculum was obtained by a 1 : 5 dilution made from 0.5 standard McFarland suspension. Solid media inocula were obtained by 10−2, 10−3, 10−4 and 10−5 dilutions made from 1 McFarland.
MGIT AST was performed according to the manufacturer's recommendations. In brief, lyophilized streptomycin, isoniazid, rifampicin and ethambutol (BBL MGIT AST SIRE kit) were reconstituted with sterile distilled water, and 0.1 mL of suspension was added to the four MGITs to give the following final concentrations: streptomycin 0.8 mg/L, isoniazid 0.1 mg/L, rifampicin 1.0 mg/L and ethambutol 3.5 mg/L. These drug-containing MGITs and one drug-free MGIT (GC) were supplemented with 0.5 mL of oleic acid–albumin–dextrose–catalase (OADC BBL MGIT) prior to inoculation with 0.5 mL of the appropriate dilution of test culture and incubated at 37 °C for up to 12 days. To test bacterial contamination, a trypticase–soy agar blood plate (BBL) was inoculated with 0.5 mL of MGIT AST inoculum and incubated for up to 3 days. A fluorescent positive control was prepared by removing the broth and replacing it with 0.4% (w/v) sodium sulfite solution, and an uninoculated MGIT, which shows no fluorescence, was used as negative control.
Susceptibility testing on solid media was performed as described previously [8–11], Middlebrook 7H10 agar plates were inoculated with 100 µL and LJ with 200 µL of the inoculum. These media were incubated at 37 °C in an atmosphere of 6% CO2 for up to 3 weeks. Solid media were inspected twice-weekly, beginning on the second week of incubation.
MGITs AST were inspected daily, beginning on the third day after inoculation. Once the MGIT GC was positive (fluoresces), the drug-containing MGITs AST were interpreted on the same day or for up to 2 additional days. The organism was then considered as resistant when the drug-containing MGIT AST was positive and susceptible when it was negative within these 2 days. The test was invalidated when the MGIT AST GC did not fluoresce until the twelfth day of incubation.
The following performance characteristics were calculated as described previously [2,3,12]: sensitivity (ST—ability to detect true resistance), specificity (SP—ability to detect true susceptibility), predictive value of sensitivity (PVS—ratio of true susceptibility to total susceptibility), predictive value of resistance (PVR—ratio of true resistance to total resistance) and efficiency (EF—ratio between the number of concordant results and the total number of results). Statistical analyses were performed with EPI Info version 6.1.
Of the 71 isolates tested, results were available for 70; one isolate did not show growth up to 12 days. The turnaround time for drug susceptibility testing for MGIT AST was 6.2 days, ranging from 5 to 10 days, while those for 7H10 MOP and LJ MOP were 18 and 21 days, respectively. The QC and reproducibility performed with control strains gave the same pattern.
As summarized in Tables 1 and 2, all three methods agreed for 61 isolates; 57 were susceptible and four resistant to at least one drug. MGIT AST agreed with 7H10 for 61 isolates (EF 87.1%); 57 were susceptible and four resistant to at least one drug. LJ agreed with 7H10 for 63 isolates (EF 90%); 59 were susceptible and four resistant to at least one drug.
Table 1. Analysis of susceptibilities by the three methods
Table 2. Analyses of susceptibilities by MGIT AST and LJ MOP
The results of the three methods agreed for 64 isolates; 61 were susceptible and three resistant. MGIT AST and 7H10 MOP were concordant for 64 isolates (EF 91.4%); 61 isolates were susceptible and three resistant. Of the six discrepant isolates, five were resistant by 7H10 MOP and susceptible by MGIT AST, and one was resistant by MGIT AST and susceptible by 7H10 MOP. LJ and 7H10 MOP were concordant for 65 isolates (EF 92.2%); 62 were susceptible and three resistant. The five discrepant isolates were resistant by 7H10 MOP and susceptible by MOP. No statistically significant difference was seen between the two combinations (P = 0.62). MGIT AST versus LJ MOP showed only one discrepancy, in which the isolate was resistant by MGIT and susceptible by LJ.
All three methods agreed for 68 isolates (EF 97.1%); 62 were susceptible and six resistant. The two discrepant isolates were resistant by MGIT AST and susceptible by 7H10 MOP. Both MOP agreed for all isolates; 64 were susceptible and six resistant. No statistically significant difference was observed between MGIT AST and LJ MOP (P = 0.46).
The results of the three methods agreed for all isolates; 66 were susceptible and four resistant.
All three methods agreed for 66 isolates; 65 were susceptible and one resistant. MGIT AST and 7H10 MOP were concordant for 66 isolates (EF 94.2%); 65 were susceptible and one resistant. Of the four discrepant isolates observed, three were resistant by 7H10 MOP and susceptible by MGIT AST, while one was resistant by MGIT AST and susceptible by 7H10 MOP. Both MOP were concordant for 67 isolates (95.7%); 66 were susceptible and one resistant. The three discrepant isolates were resistant by the reference method and susceptible by LJ MOP (P = 0.93). MGIT AST versus LJ MOP showed one discrepancy, in which the isolate was resistant by MGIT AST and susceptible by LJ MOP.
For the overall discrepancies observed between MGIT AST/LJMOP and the reference method, most of the discrepant isolates were susceptible by MGIT AST/LJ MOP and resistant by the reference method.
Details of these discrepancies are summarized in Tables 1 and 2. Overall, in terms of results, MGIT AST was closer to LJ MOP than to 7H10 MOP.
The MGIT AST performance characteristics summarized in Table 3 show ST, SP, PVR and PVS of 100% for rifampicin, while those of streptomycin, isoniazid and ethambutol ranged from 25% to 100%.
Table 3. Performance characteristics for MGIT AST and LJ MOP as compared to 7H10
| MGIT AST||37.5||98.4||92.5||75||91.4|
| MGIT AST||100||96.8||100||75||97.1|
| MGIT AST||100||100||100||100||100|
| MGIT AST||25||98.4||95.5||50||94.2|
The increase in mycobacterial disease has stimulated the need to develop more rapid and reliable methods for diagnostic and susceptibility testing. It is widely known that drug susceptibility testing of M. tuberculosis is more rapid in liquid medium than in solid medium [3,4,13]. Among liquid media used for susceptibility testing of M tuberculosis, the BACTEC System is the only commonly used method in public health, despite its cost and the radioactive material required. The recently introduced MGIT AST has been reported to have good reliability for susceptibility testing of M. tuberculosis[4,6].
This study was performed to evaluate the reliability of MGIT AST as compared to LJ MOP for the first-line drugs streptomycin, isoniazid, rifampicin and ethambutol, using 7H10 MOP as the reference method.
The turnaround time for MGIT AST in this study was 6.2 days, compared to 21 days for the MOP. Similar data have been reported previously [6,7,14]. This short time to complete susceptibility testing combined with the short time to detection of M. tuberculosis in clinical specimens reported earlier  could allow laboratories using this medium to reduce the delay in completing the mycobacteria tests, as suggested by the CDC .
Among the 70 isolates, MGIT AST agreed with 7H10 MOP and LJ MOP for 87.1% and 95.7% of isolates, respectively. Results obtained with MGIT AST are divided into the following groups:
- 1Isoniazid and rifampicin with 97.1% and 100% efficiency, respectively. Only two discrepancies were observed between MGIT AST and the two MOP; the discrepant isolates were resistant by MGIT AST and susceptible by both MOP. This excellent agreement confirmed findings reported previously [4–7,15–16].
- 2Streptomycin and ethambutol with 91.4% and 94.2% efficiency, respectively. There were six and four discrepancies between MGIT AST and 7H10 MOP, respectively, for streptomycin and ethambutol. Similar findings were reported previously [5,17]. In this study, and that performed by Gerome et al , most of the discrepant isolates observed with streptomycin and ethambutol were false positives. The same findings were reported by other studies [5,17]. This number of false positives with streptomycin and ethambutol detected by MGIT AST seems more likely due to the inability of MGIT AST to detect low levels of resistance (low proportion of resistant bacilli). In BACTEC 460 this problem was solved by comparing test vials with a 1 : 100 diluted GC.
MGIT AST performance characteristics were high for isoniazid and rifampicin, except for isoniazid PVR. These performance characteristics were similar to those reported earlier [3,4]. On the other hand, streptomycin and ethambutol showed low ST and PVR, while SP and PVS were high. This difference in values was due to the number of false negatives detected by MGIT AST. In terms of efficiency, when compared to LJ MOP, the MGIT AST reached an agreement level of 95% (97.1–100%). Similar results were observed by Lazlo et al  when comparing BACTEC 460 with LJ.
In summary, our data suggest the following conclusions: (1) in terms of time, MGIT AST is a time-saving method and can be used as an alternative to the BACTEC System to shorten the delay in reporting the susceptibility results; (2) in terms of accuracy, MGIT AST is a reliable method for the susceptibility testing of M. tuberculosis complex with rifampicin and isoniazid and can be used for the rapid screening of multiresistant strains. Further studies need to be performed including a high number of resistant strains for streptomycin and ethambutol to better evaluate their sensitivity and PVR. For all the first-line drugs, MGIT AST can be used in place of LJ MOP as a susceptibility testing method.
Our thanks go to Salman H. Siddiqi for helpful suggestions, and to Becton Dickinson for providing us with MGIT AST reagents.