Evaluation of the BD Bactec MGIT 320 System for Detection of Mycobacteria and Drug Susceptibility Testing of Mycobacterium tuberculosis

Abstract

Two different laboratories evaluated growth and detection of mycobacteria and drug susceptibility testing of Mycobacterium tuberculosis by the BD Bactec MGIT 320 using the BD Bactec MGIT 960 (BD Diagnostics, Sparks, MD) as a reference method. Out of 359 processed sputum specimens for detection of mycobacteria, 99.7% were in agreement between the MGIT 320 and MGIT 960. Streptomycin (STR), isoniazid (INH), rifampin (RIF), ethambutol (EMB) (collectively known as SIRE), and pyrazinamide (PZA) drug susceptibility testing was performed on 89 clinical strains, prepared from both liquid and solid inocula. The results of SIRE and PZA were 100% reproducible between the two instruments tested at both laboratories.

TEXT

Tuberculosis (TB) is a serious global public health problem. Twelve million prevalent cases were reported worldwide in 2010 (5). To achieve early diagnosis and effective treatment of TB, rapid and accurate drug susceptibility testing (DST) methods must be used. The World Health Organization and the U.S. Centers for Disease Control and Prevention have recommended the use of liquid culture systems for DST and for improving time to detection (1, 2, 3). DST of Mycobacterium tuberculosis (MTB) with the BD Bactec MGIT 960 system produces accurate results more rapidly than the conventional agar proportion method (4, 5, 7). The BD Bactec MGIT 320 contains only one drawer, holds 320 MGIT tubes, can be placed on a benchtop, and uses the same technology as the MGIT 960. It is designed for laboratories with limited space and workload. For example, MGIT 320 could be used by a laboratory that cultures up to 40 mycobacteriology specimens per week, with an incubation period of 6 weeks, and performs up to five drug susceptibility tests per week. From April 2010 to February 2011, a study was conducted by the Microbial Diseases Laboratory (MDL) of the California Department of Public Health (CDPH) and the National Tuberculosis Reference Laboratory (NTRL) of Bangkok, Thailand, to evaluate the performance of MGIT 320 for detection of mycobacteria and DST for streptomycin, isoniazid, rifampin, ethambutol (collectively known as SIRE), and pyrazinamide (PZA). Here we report the results obtained using the MGIT 960 system as a reference.

This study compared the detection of growth of mycobacteria and DST performance of the MGIT 320 with the MGIT 960. Sputum samples were processed by the N-acetyl-l-cysteine (NALC)-NaOH method (3) and inoculated in parallel for growth detection in both MGIT instruments and also inoculated on solid media. The growth in instrument-positive MGIT tubes was stained for acid-fast bacilli (AFB) and identified. Identification of Mycobacterium tuberculosis complex was accomplished at MDL by use of a molecular beacon assay (6) and at the NTRL by the presence of characteristic cording morphology, growth rate, and lack of pigmentation.

Isolates of various drug susceptibility profiles were chosen and prepared in parallel for DST. DST inocula were prepared from both solid (Lowenstein Jensen [LJ]) and liquid (MGIT) media following the manufacturer's protocol. The SIRE and PZA results obtained with the MGIT 320 system were compared with those obtained with MGIT 960. Any assays yielding discrepancies were repeated once for final data analysis.

For growth detection of Mycobacterium species, the NTRL tested 359 sputum specimens. For DST, the two laboratories tested a total of 89 strains, including 35 pansusceptible strains and 54 resistant strains with various patterns of drug resistance. The NTRL tested 33 strains using MGIT as the source of inoculum and 30 strains using LJ as the source of inoculum. The MDL tested 27 strains in parallel, one set using LJ and the other set using MGIT as the source of inoculum.

For growth detection, 0.8 ml of BD PANTA antibiotic mixture was added to each MGIT tube to be inoculated with 0.5 ml of a processed sputum. For DST preparation, 0.8 ml of growth supplement provided in the SIRE or PZA kits and 0.1 ml of the specific drug were added to each MGIT or PZA medium tube. Each isolate was prepared in DST sets of 5 tubes (control and SIRE) and 2 tubes (control and PZA). The control tube contained growth supplement without drug. The test concentrations for SIRE were 1.0 μg/ml STR, 0.1 μg/ml INH, 1.0 μg/ml RIF, and 5.0 μg/ml EMB, and that of PZA was 100 μg/ml.

The 359 sputum specimens were processed using the NALC-NaOH decontamination/digestion method. After specimen processing, 0.5 ml of the concentrated sputum sediment was added to two MGIT tubes and one drop to solid medium. One tube was placed in the MGIT 320 and the other in the MGIT 960 system. The solid medium was incubated at 35 to 38°C.

Inocula for DST were prepared from LJ cultures within 14 days of growth detection and MGIT cultures after 1 to 5 days of incubation following detection of growth. SIRE and PZA were set up in parallel sets following the BD package insert protocols. One set was placed in the MGIT 320, and the other was placed simultaneously in MGIT 960.

M. tuberculosis H37Rv (ATCC 27294) was set up once every week for SIRE and PZA as a quality control (QC) strain using the same procedure as the test strains. If the QC strain did not yield pansusceptible results for SIRE and PZA, strains tested the same week as the failed QC results were tested again. Tests producing any discrepancies in DST results between the two instruments or yielding error codes were repeated once.

NTRL.

Table 1 shows NTRL's comparison of the MGIT 320 and MGIT 960 systems for SIRE DST sets prepared from liquid MGIT medium, and Table 2 shows the comparison with inocula prepared from solid LJ medium. A total of 564 DST results were generated from 282 parallel sets of cultures between the two instruments. All results (from all 282 parallel sets) were in 100% agreement.

Table 1.

NTRL comparison of MGIT 960 and MGIT 320 DST results (MGIT inoculum)

Drug	MGIT 320 result	No. of samples with MGIT 960 result		Overall % agreement
		S	R
STR	S	26	0	100 (33/33)
	R	0	7
INH	S	24	0	100 (33/33)
	R	0	9
RIF	S	30	0	100 (33/33)
	R	0	3
EMB	S	32	0	100 (33/33)
	R	0	1

Table 2.

NTRL comparison of MGIT 960 and MGIT 320 DST results (LJ inoculum)

Drug	MGIT 320 result	No. of samples with MGIT 960 result		Overall % agreement
		S	R
STR	S	25	0	100 (30/30)
	R	0	5
INH	S	24	0	100 (30/30)
	R	0	6
RIF	S	28	0	100 (30/30)
	R	0	2
EMB	S	29	0	100 (30/30)
	R	0	1
PZA	S	30	0	100 (30/30)
	R	0	0

Table 3 shows the growth agreement between the MGIT 320 and MGIT 960 systems. Out of 359 processed sputum sediments, 216 tested as true negatives and 134 as true mycobacterium positives. One discrepancy occurred; a sample tested negative in MGIT 320 and was identified as nontuberculosis Mycobacterium (NTM) by MGIT 960. Eight sputum specimens were contaminated with nonmycobacterial contaminants, including three MGIT 320 and five MGIT 960 cultures. The contamination rates in MGIT 320 (0.8%) and in MGIT 960 (1.4%) were similar, and the difference was not considered to be due to instrumentation. Three out of the four NTMs isolated were identified as group III nonphotochromogens, and one was a group II scotochromogen. Overall, growth data obtained with the MGIT 320 and MGIT 960 were in 99.7% agreement.

Table 3.

NTRL agreement between MGIT 320 and MGIT 960 growth results

MGIT 320 result	No. of samples with MGIT 960 result
	No growth	Growth
No growth	216	1
Growth	0	134

MDL.

Table 4 shows MDL's comparison data. Each system produced 135 DST results. IRE and PZA were all in 100% agreement between the two instruments as well as the archived results. A discrepancy occurred between streptomycin test for the same isolate, inoculated from different media. The strain tested STR susceptible by MGIT 960 but resistant by MGIT 320, with growth units of 374 and 121 for inoculum prepared from positive MGIT and LJ, respectively. After retesting, the strain tested susceptible by both MGIT 960 and MGIT 320, which was concordant with the archived results. A 100% correlation was realized for SIRE and PZA between MGIT 320 and MGIT 960 after discrepancy resolution (or 99.3% correlation before discrepancy resolution [268/270 total]).

Table 4.

MDL comparison of MGIT 960 and MGIT 320 DST results^a

Drug	MGIT 320 result	No. of samples with MGIT 960 result		Overall % agreement
		S	R
STR	S	21	0	96.3 (26/27)b
	R	1	5
INH	S	18	0	100 (27/27)
	R	0	9
RIF	S	22	0	100 (27/27)
	R	0	5
EMB	S	24	0	100 (27/27)
	R	0	3
PZA	S	25	0	100 (27/27)
	R	0	2

^aResults were the same for MGIT and LJ media.

^b100% agreement when retested.

The correlation of mycobacterial detection from processed sputum and DST SIRE and PZA results between the MGIT 320 and MGIT 960 using both types of media was excellent. Although streptomycin resulted in a single discrepant strain, 100% agreement was realized after retesting. The isolate that initially tested STR resistant may have possibly been a borderline STR-resistant strain, suggested by the fact that the number of growth units in the STR drug tubes was under 400 (374 and 121). Overall, reproducibility of DST results was nearly perfect for all five drugs tested. The results from this study verify the equivalency between the MGIT 320 and MGIT 960. Note that while the instrument cost and space requirements for the MGIT 320 will be less than those of the MGIT 960, some of the same constraints will apply in resource-limited settings, namely, the cost of MGIT media and supplements and the need for a containment laboratory with well-trained personnel and rigorous safety precautions.