N. Banaiee et al., Luciferase reporter mycobacteriophages for detection, identification, and antibiotic susceptibility testing of Mycobacterium tuberculosis in Mexico, J CLIN MICR, 39(11), 2001, pp. 3883-3888
The utility of luciferase reporter mycobacteriophages (LRPs) for detection,
identification, and antibiotic susceptibility testing of Mycobacterium tub
erculosis was prospectively evaluated in a clinical microbiology laboratory
in Mexico City, Mexico. Five hundred twenty-three consecutive sputum sampl
es submitted to the laboratory during a 5-month period were included in thi
s study. These specimens were cultivated in Middle-brook 7H9 (MADC), MGIT,
and Lowenstein-Jensen (LJ) media. Of the 71 mycobacterial isolates recovere
d with any of the three media, 76% were detected with the LRPs, 97% were de
tected with the MGIT 960 method, and 90% were detected with LJ medium. When
contaminated specimens were excluded from the analysis, the LRPs detected
92% (54 of 59) of the cultures. The median time to detection of bacteria wa
s 7 days with both the LRPs and the MGIT 960 method. LRP detection of growt
h in the presence of p-nitro-alpha -acetylamino-beta -hydroxypropiophenone
(NAP) was used for selective identification of M. tuberculosis complex (MTC
) and compared to identification with BACTEC 460. Using the LRP NAP test, 4
7 (94%) out of 50 isolates were correctly identified as tuberculosis comple
x. The accuracy and speed of LRP antibiotic susceptibility testing with rif
ampin, streptomycin, isoniazid, and ethambutol were compared to those of th
e BACTEC 460 method, and discrepant results were checked by the conventiona
l proportion method. In total, 50 MTC isolates were tested. The overall agr
eement between the LRP and BACTEC 460 results was 98.5%. The median LRP-bas
ed susceptibility turnaround time was 2 days (range, 2 to 4 days) compared
to 10.5 days (range, 7 to 16 days) by the BACTEC 460 method. Phage resistan
ce was not detected in any of the 243 MTC isolates tested. Mycobacteriophag
e-based approaches to tuberculosis diagnostics can be implemented in clinic
al laboratories with sensitivity, specificity, and rapidity that compare fa
vorably with those of the MGIT 960 and BACTEC 460 methods. The phages curre
ntly provide the fastest phenotypic assay for susceptibility testing.