Direct Detection of Rifampin-Resistant Mycobacterium tuberculosis in ...

JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 2004, p. 4438–4443 0095-1137/04/$08.00⫹0 DOI: 10.1128/JCM.42.10.4438–4443.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Vol. 42, No. 10

Direct Detection of Rifampin-Resistant Mycobacterium tuberculosis in Respiratory Specimens by PCR-DNA Sequencing W. C. Yam,1 C. M. Tam,2 C. C. Leung,2 H. L. Tong,1 K. H. Chan,1 E. T. Y. Leung,1 K. C. Wong,1 W. W. Yew,3 W. H. Seto,1 K. Y. Yuen,1 and P. L. Ho1* Centre of Infection and Department of Microbiology, Queen Mary Hospital, The University of Hong Kong,1 Tuberculosis and Chest Service, Department of Health,2 and Tuberculosis and Chest Unit, Grantham Hospital,3 Hong Kong Special Administrative Region, People’s Republic of China Received 18 April 2004/Returned for modification 9 June 2004/Accepted 28 June 2004

This study evaluated the feasibility of a molecular strategy based on identification of Mycobacterium tuberculosis by IS6110 PCR or Cobas Amplicor PCR, and rpoB PCR-DNA sequencing of the 81-bp rifampin resistance determining region (RRDR) for direct detection of rifampin resistance in respiratory specimens. A collection of 2,138 respiratory specimens and 352 nonduplicate M. tuberculosis isolates (including 233 isolates from the evaluated respiratory specimens and an additional collection of 119 stored isolates) from Southern China was investigated. Using culture as the reference test, the overall diagnostic sensitivities of an acid-fast bacillus (AFB) smear, Cobas Amplicor PCR, IS6110 PCR were 54.5% (156 of 286), 86.7% (248 of 286), and 89.2% (255 of 286), respectively. The sensitivities of the rpoB PCR for the specimens with positive AFB smears and with positive PCR results in the IS6110 PCR and/or Cobas Amplicor PCR were 100% (156 of 156) and 92.3% (239 of 259), respectively. Of the 352 nonduplicate M. tuberculosis isolates, the agar proportion method for rifampin reported 39 resistant strains. Full agreement (352 of 352) was found with the agar proportion method and the genotype inferred from the rpoB DNA sequencing data for rifampin. Thirty-nine mutations of nine distinct kinds, eight point mutations, and one deletion within the RRDR were found in the 39 resistant strains. For the direct DNA sequencing performed on rpoB PCR-positive respiratory specimens, the concordance with the agar proportion method and the subsequent PCR-sequencing for the culture isolate was 100%. This strategy has potential application for direct and rapid diagnosis of rifampin-resistant M. tuberculosis in IS6110 PCR or Cobas Amplicor PCR-positive respiratory specimens. Tuberculosis remains an important disease worldwide. The emergence of multidrug resistance in strains of Mycobacterium tuberculosis (M. tuberculosis) is one of the most important threats to tuberculous control. Despite continued efforts directed to improve tuberculosis control programs at national and international levels, recent surveys reveal that drug-resistant tuberculosis is still ubiquitous and rates of infection are alarmingly high in several countries. According to Espinal et al. (9), multidrug-resistant tuberculosis (MDR-TB) defining as resistance to at least isoniazid (INH) and rifampin (RIF) was present among new cases in 54 of 58 surveyed geographic sites in 1996 to 1999. Among new cases, alarming rates were reported in Estonia (14.1%), Latvia (9.0%), Tomsk Oblast of Russian Federation (6.5%), Ivanovo Oblast of the Russian Federation (9.0%), and Henan of China (10.8%). Overall, in all 58 surveyed sites and among the previously treated cases (9), the median MDR-TB rate was 9.3%, and for 35 geographic sites it was present at rates of 5% or higher. In M. tuberculosis, monoresistance to RIF is rare, and at least 90% of all RIF-resistant clinical isolates are also resistant to isoniazid (7). Hence, a positive result for RIF resistance would be useful as a strong surrogate of MDR-TB. In resistant isolates, it has been shown that up to 95 to 98% RIF resistance

is caused by mutations in the rpoB gene encoding the RNA polymerase ␤-subunit. In the great majority of the RIF-resistant isolates, mutations occurred within a 81-bp hotspot region (the rifampin resistance determining region [RRDR], encoding 27 amino acids and corresponding to codons 507 to 533 or cluster I according to Escherichia coli numbering) in the center of the 3,516-bp rpoB gene (24). Within the RRDR, more than 80 distinct variants of single-base mutations or in-frame deletions or insertions have been reported among RIF-resistant isolates. Although several different molecular techniques such as heteroduplex analysis, line probe assay, and PCR singlestrand conformational length polymorphism have been used for analysis of rpoB gene mutations, these techniques were limited in that only the most frequent types of mutations were included. In the present study we report a rapid assay based on PCR and sequencing of the rpoB gene for direct detection of RRDR-associated RIF-resistant M. tuberculosis in clinical isolates and respiratory specimens. A total of 2,138 respiratory specimens and 142 M. tuberculosis isolates were used to evaluate the specificity and sensitivity of this assay. The results were compared to commercial and in-house diagnostic assays for M. tuberculosis and phenotypic drug susceptibility testing.

* Corresponding author. Mailing address: Department of Microbiology, University Pathology Building, Queen Mary Hospital, Pokfulam Rd., Hong Kong Special Administrative Region, People’s Republic of China. Phone: 852-28554892. Fax: 852-28551241. E-mail: plho@hkucc .hku.hk.

Specimens and isolates. Between March 1999 and October 2002, 2,138 respiratory specimens (1,807 expectorated sputum and 331 bronchoalveolar lavage samples) were collected from 1571 patients suffering from chest symptoms and/or chest radiographic infiltrates of undetermined origin, including 712 pa-

MATERIALS AND METHODS

4438

DIRECT DETECTION OF RIF-RESISTANT M. TUBERCULOSIS

VOL. 42, 2004

4439

TABLE 1. Comparative sensitivities and specificities of three PCR assays for detection of M. tuberculosis No. of positive results/total no. (%)a Sample (n)

Respiratory specimens (2,138)b MTB culture positive (286) Smear positive (156) Smear negative (130) NTM culture positive (66)c Culture negative for Mycobacterium spp. (1,786) Purified M. tuberculosis isolates (119)

rpoB PCR

IS6110 PCR (UE)

Cobas Amplicor PCR (UE)

UE

3⫻E

156/156 (100) 99/130 (76) 0/66 (0) 4/1,786 (0.2)e

156/156 (100) 92/130 (71) 0/66 (0) 4/1,786 (0.2)e

128/156 (82) 54/130 (42) 6/66 (9%)d 4/1,786 (0.2)e

156/156 (100) 79/130 (61) 6/66 (9%)d 4/1,786 (0.2)e

ND

ND

119/119 (100)

ND

a

UE, uncontentrated extract; 3⫻E, threefold-concentrated extract. ND, not done. The 2,138 respiratory specimens were obtained from 1,571 patients. Overall, 286 specimens from 233 patients were culture positive for M. tuberculosis. NTM, nontuberculous mycobacteria, including M. avium-M. intracellulare complex (n ⫽ 25), M. chelonei (n ⫽ 16), M. gordonae (n ⫽ 8), M. scrofulaceum (n ⫽ 8), and M. fortuitum (n ⫽ 9). d The six rpoB PCR products were sequenced and found to align with M. avium-intracellulare complex (n ⫽ 5) and M. gordonae (n ⫽ 1) sequences (accession AY544889 and AY544919). e The four culture-negative samples were obtained from four different patients with tuberculosis as confirmed by positive-culture results from previous respiratory or concurrent extrapulmonary specimens. b c

tients in Queen Mary Hospital, 642 patients in Grantham Hospital, and 217 outpatients in Polyclinics of the Department of Health in Hong Kong. An additional 119 nonduplicate clinical isolates of M. tuberculosis were collected from three major cities in Pearl River Delta of the South China region of the Peoples Republic of China: Hong Kong (84 isolates), Macau (23 isolates), and Guangzhou (12 isolates) (16). After a direct smear for acid-fast bacilli (AFB), respiratory specimens were decontaminated by using N-acetyl-L-cysteine-sodium hydroxide and concentrated by centrifugation as described previously (21). The digested sediments were divided equally for AFB culture and subsequent PCR assays. Cultures positive for AFB were identified by using the AccuProbe hybridization assay (Gen-Probe, Inc., San Diego, Calif.) and conventional biochemical tests. Antimicrobial susceptibility testing for INH, RIF, streptomycin (STR), pyrazinamide (PZA), and ethambutol (EMB) was performed by using the agar proportion method (20). In brief, growth from the primary isolation medium was subcultured onto Lowenstein-Jensen (LJ) medium. The bacterial inoculum was prepared by picking fresh colonies from the LJ medium and adjusted to a turbidity matching that for a McFarland 1 standard. A 0.1-ml aliquot of 100- and 10,000-fold dilution samples were placed onto drug-containing and no-drug sectors of Middlebrook 7H10 quadrant plates. The following drug concentrations were used: INH (0.2 ␮g/ml), RIF (1 ␮g/ml), PZA (25 ␮g/ml), EMB (5 ␮g/ml), and STR (2 ␮g/ml). Inoculated plates were examined weekly for up to 3 weeks. Once the no-drug control quadrant of either dilution shows at least 50 to 150 colonies, the plates were read and interpreted. If the percentage of the colonies in the drug-containing sector is ⱖ1%, the result is interpreted as resistant. In addition to mycobacterial culture, conventional bacterial cultures were set up by using the following media: 5% defibrinated horse blood agar, chocolate agar, and MacConkey agar. The inoculated plates were incubated at 37°C in 5% CO2 and were examined after overnight incubation. The following bacteria were identified by Gram stain, colony morphology, and conventional biochemical tests (1): Streptococcus pneumoniae, Haemophilus influenzae, Staphylococcus aureus, Enterobacteriaceae, and nonfermentors. DNA extraction for PCR. By using the Roche Cobas Amplicor extraction kit, 500 ␮l of wash solution was added to 100 ␮l of the decontaminated sediment (33, 36). After centrifugation at 12,500 ⫻ g for 10 min and decanting, 100 ␮l of lysis reagent was added, and the mixture was vortex mixed, followed by incubation at 60°C for 45 min. Extraction was terminated by adding 100 ␮l of neutralization reagent, and the resultant mixture was used as DNA template (undiluted extract). As an additional step, 100 ␮l of the undiluted extract was further cleaned up by using minicolumns (gel extraction kit; Qiagen, Hilden, Germany) according to the manufacturer’s instructions, and then 30 ␮l of purified DNA was eluted (threefold-concentrated extract). For purified culture of M. tuberculosis on LJ medium, a uniform bacterial suspension (McFarland standard no.1) was made in 0.1 M Tris-HCl (pH 7.5). For the M. tuberculosis standard strain H37Rv, the bacterial suspension was further diluted from 10⫺3 to 10⫺10 for colony counting by using Middlebrook agar plates. All bacterial suspensions were subsequently heated at 100°C for 30 min, frozen at ⫺80°C, thawed at 80°C, and centrifuged at 15,000 ⫻ g for 10 min; the supernatant was used directly for subsequent PCR assays. For the M. tuberculosis

standard strain H37Rv, serial dilutions of heated extract were used for determination of analytical sensitivity of IS6110, 16S rRNA, and rpoB PCR assays. PCR for M. tuberculosis 16SrRNA and IS6110. The Cobas Amplicor M. tuberculosis test (Roche) uses Mycobacterium genus-specific biotinylated primers to amplify a sequence of 584-bp within a 1,500-bp region encoding the 16S rRNA of M. tuberculosis (36). An internal control was incorporated in each reaction for monitoring of PCR inhibitors. A manual one-tube nested PCR for IS6110 was performed as described previously (3, 33, 36). A 50-␮l and a 10-␮l aliquot of the undiluted DNA extract were used as the PCR templates in the Cobas Amplicor test and IS6110 assay, respectively. PCR and DNA sequencing for rpoB. Each PCR contained 10 ␮l of DNA template (undiluted or threefold-concentrated extract). For amplification of rpoB gene, PCR was performed in a final volume of 50 ␮l containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl (Applied Biosystems); 1.25 mM MgCl2 (Applied Biosystems); 200 ␮M (each) deoxynucleoside triphosphate (Cobas Amplicor Diagnostics); 0.4 ␮M (each) primers (Life Technologies), and 2 U of AmpliTaq Gold Polymerase (Applied Biosystems). The primers used to amplify the rpoB gene fragment were modified from the TR8 and TR9 primer pair reported by Telenti et al. (28): forward primer TR8x (5⬘-TCGCCGCGATCAAGGAGTTC TTCGGC-3⬘; positions 76786 to 76811 under accession no. BX842574 [underlined bases represent the original primer sequence]) and reverse primer TR9x (5⬘-TGCACGTCGCGGACCTCCAGCCCGGCAC-3⬘; positions 76942 to 76915). AmpliTaq Gold polymerase in the master mix was first activated by incubation at 94°C for 12 min. The reaction mixture was then subjected to 40 cycles of amplification (denaturation at 94°C for 1 min and annealing and extension at 72°C for 2 min), followed by a final 7-min extension at 72°C. After PCR amplification, a 10-␮l aliquot of the PCR product was electrophoresed for 1 h through 2% agarose gel, and the target band of 157-bp was visualized under UV illumination. This 157-bp DNA fragment included the complete cluster I region (positions 507 to 533) of rpoB gene. For positive samples, the 157-bp fragment was subjected to cycle sequencing by the ABI Prism dye terminator cycle sequencing ready reaction kit (version 2.0) at a quarter of the recommended reaction volume and an ABI 377 genetic analyzer (Applied Biosystems). For all samples, the sequences of both strands of the amplicons were determined. The generated sequences were assembled and edited by using EDITSEQ v.4.0 program in DNASTAR software (Lasegene). The edited sequence was further compared to published sequence of rpoB gene for H37Rv (GenBank accession BX842574) to identify RIF resistance in terms of any mutations, insertions, and deletions in RRDR (codons 507 to 533). The nucleotide and amino acid numbering was based on homologous mutations in E. coli (29).

RESULTS Sensitivity and specificity of the PCR assays. Of the 2,138 respiratory specimens shown in Table 1, 286 (13.4%) specimens from 233 patients were culture positive for M. tubercu-

4440

YAM ET AL.

FIG. 1. Comparison of two primer pairs for direct PCR amplification of rpoB in M. tuberculosis in respiratory specimens. PCR with primers TR8 and TR9 (28) exhibited nonspecific amplification (lanes 1 to 5), whereas a modified pair of primers, TR8x and TR9x, produced highly specific 157-bp rpoB amplicons (lanes 6 to 10). The respiratory specimens tested in lanes 1 to 5 and lanes 6 to 10 were the same. Results for samples found to be culture negative for M. tuberculosis (lanes 1 and 6) and culture positive for M. tuberculosis (lanes 2 to 5 and lanes 7 to 10) are shown. M, size marker. See Materials and Methods for the primer sequences.

losis and 66 (3.1%) were culture positive for nontuberculous mycobacteria (NTM), while the remaining 1,786 (83.5%) were culture negative for Mycobacterium. By using culture as the “gold standard,” the sensitivity of the AFB smear method for the detection of M. tuberculosis was 54.5%. In respiratory specimens, the overall diagnostic sensitivities of the IS6110 PCR, Cobas Amplicor PCR and rpoB PCR assays for the unconcentrated extract were 89.2, 86.7, and 63.6%, respectively. After concentration of DNA template, all 156 (100%) AFB smearpositive specimens and 79 (61%) AFB smear-negative specimens were positive in the rpoB PCR, thus increasing the overall rpoB PCR sensitivity to 82.2%. The sensitivities of the rpoB PCR for the specimens with a positive AFB smear and a positive PCR result in the IS6110 PCR and/or Cobas Amplicor PCR analyses were 100% (156 of 156) and 92.3% (239 of 259), respectively. For the purified M. tuberculosis isolates, all of the 119 isolates could be amplified in the rpoB PCR. With regard to the analytical sensitivities, endpoint detection of M. tuberculosis for IS6110 PCR and Cobas Amplicor PCR was found to be 1.5 CFU, whereas that of rpoB PCR was found to be 15 CFU. This might be explained by that H37Rv has 17 copies of IS6110 per genome. A single 157-bp band corresponding to the rpoB amplicons or no amplification was obtained in all of the 2,138 respiratory specimens. In contrast, nonspecific amplification was found for the original primer pair used by Telenti et al. (28) (Fig. 1). In conventional bacterial culture, 135 of the 2,138 respiratory specimens were positive for the common nonmycobacterial pathogens, including Streptococcus pneumoniae (n ⫽ 36), H. influenzae (n ⫽ 41), Pseudomonas aeruginosa (n ⫽ 24), Staphylococcus aureus (n ⫽ 21), and Acinetobacter spp. (n ⫽ 13). In all of the 135 specimens, no false-positive PCR results were

J. CLIN. MICROBIOL.

obtained. Discrepant results between the rpoB PCR results and M. tuberculosis culture results were found for 10 specimens (Table 1), including five samples culture positive for M. aviumM. intracellulare complex, one sample culture positive for M. gordonae, and four samples culture negative for Mycobacterium. For the four Mycobacterium culture-negative samples, the sequenced rpoB PCR products had 100% identity to M. tuberculosis. The four samples were obtained from different patients with tuberculosis confirmed by positive culture from previous respiratory or concurrent extrapulmonary specimens. For the six NTM culture-positive samples, DNA sequencing confirmed false amplification of the rpoB gene from the NTM species. Therefore, the specificity of the rpoB PCR assay was 99.7% (1,782 of 1,788). First-line drug resistance patterns. The susceptibilities of the 352 M. tuberculosis isolates (233 isolates from respiratory specimens and 119 isolates from purified culture) to INH, RIF, STR, PZA, and EMB were determined by the agar proportion method. Overall, 88.9% (313 of 352) of the isolates were susceptible to rifampin and 39 were resistant. Only five showed resistance to only RIF, whereas 34 isolates were at least resistant to INH and RIF. Of these MDR-TB isolates, 29 strains showed further resistance to at least one other first-line drug and two were resistant to RIF, INH, STR, PZA, and EMB. The number of RIF-resistant M. tuberculosis strains from Hong Kong, Macau, and Guangzhou were 21, 12, and 6, respectively. Correlation between susceptibility testing and rpoB mutations. The 157-bp amplicons from the 239 rpoB PCR-positive respiratory samples and the 428 purified M. tuberculosis isolates were analyzed by direct sequencing. A comparison of the genotypic prediction with susceptibility results obtained by the agar proportion method is shown in Table 2. For the 235 rpoB PCR-positive and M. tuberculosis culture-positive respiratory samples, the rpoB DNA sequences were identical to those for the corresponding M. tuberculosis isolates. For the four rpoB PCR-positive and M. tuberculosis culture-negative respiratory samples, no rpoB mutations were found. For all 313 RIF-

TABLE 2. Comparison of genotypic testing by rpoB PCR-sequencing to the agar proportion method for 286 M. tuberculosis culture-positive respiratory specimens and 352 purified isolates

Sample for genotypic testing and result (n)

No. of isolates (no. of individual patient strainsa) with the indicated result as determined by the agar proportion method Susceptible

Resistant

M. tuberculosis culture-positive respiratory specimens (286)a RRDR mutation absent (230) RRDR mutation present (5) PCR negative (51)

230 (184) 0 51 (45)

0 5 (4) 0

Purified nonduplicate isolates (352) RRDR mutation absent (313) RRDR mutation present (39)

313 0

0 39b

a The 286 culture-positive specimens were obtained from 233 patients. The numbers of individual patient strains, as defined by a unique genotype, are shown in parentheses. b The 39 RIF-resistant strains included four from the tested respiratory specimens and 35 from the 119 purified isolates.


VOL. 42, 2004

susceptible isolates, there was no mutation in the RRDR. In contrast, RRDR mutations were found for all 39 RIF-resistant isolates. One had a double mutation (Gln513Arg and His526Asn). Of the remaining 38 mutations, one was a deletion (518Asn) and the remaining 37 had missense mutations which led to the following amino acid substitutions: Ser531Leu (n ⫽ 15), Ser531Tyr (n ⫽ 7), His526Arg (n ⫽ 5), His526Asp (n ⫽ 3), Asp516Val (n ⫽ 6), and Ser522Leu (n ⫽ 1). DISCUSSION In clinical laboratories, direct detection of M. tuberculosis in clinical specimens by PCR has become increasingly popular. The IS6110 PCR and Cobas Amplicor PCR represent examples that have been extensively investigated in terms of their sensitivities and specificities. In accordance with previous studies (32, 36), we found here that both assays were 100% specific and had overall diagnostic sensitivities of ca. 90% for culturepositive respiratory specimens. Since 1999, one of the PCR assays has become routinely available for all of the regional and chest hospitals in Hong Kong. In our hands, the diagnostic sensitivities of these PCR assays were consistently higher than the 40 to 60% for AFB smear, although their sensitivities for smear-negative specimens are still limited at 60 to 70% (3, 33, 36), which is similar to those reported by other investigators (2, 17). In a cohort of 155 patients being investigated for tuberculosis, our team recently showed that the molecular diagnostic service is useful for making clinical decisions on the initiation of antituberculosis treatment, particularly among patients with equivocal presentations (5). The present study extends these findings by evaluating the direct application of a PCR-sequencing assay in clinical specimens for genotypic prediction of RIF resistance. In combination with PCR assays for M. tuberculosis identification, our rpoB PCR was 100% specific. There was no false amplification of the rpoB for a variety of nonmycobacterial pathogens. Owing to the high level of rpoB sequence homology between M. tuberculosis and some NTM, the rpoB PCR was positive for six samples containing NTM but not M. tuberculosis. In a heminested rpoB PCR with sputum samples and heteroduplex analysis, Williams et al. have also reported false-positive result due to samples containing M. avium (31). The present investigation demonstrates that this potential confusion can be readily resolved by an analysis of the rpoB sequences and a negative result in the IS6110 or Cobas Amplicor PCR. However, it should be noted that some populations have a significant number of IS6110-negative strains. In these populations, IS6110 amplification should not be relied on too much. Our 100% sensitivity for smear-positive respiratory specimens was better than the 82.8% (19) and 97.5% (25) sensitivities recently reported by others. In addition, our findings further extend this application to AFB-smear-negative respiratory specimens for both identification and rifampin resistance determination with a short turnaround time, although the sensitivity is limited at ca. 60%. As with other genotypic resistance testing, a positive result is more useful than a negative result. In M. tuberculosis, mutations in the RRDR appear to be strongly if not always predictive of RIF resistance in susceptibility testing. DNA sequencing is most likely to yield the maximal benefit obtainable for a positive genotypic test result because the full spectrum of mu-

4441

tations can be accurately identified. This is particularly relevant in the prediction of RIF resistance since more than 80 distinct allelic variants, including missense mutations, either deletions or insertions alone or in combination, have been reported in the 81-bp core RRDR (19, 24, 26, 27, 30). All positive results in our genotypic assay were “true resistance” compared to concordances of 90% for the line probe assay (6), 87.2% for amplification refractory mutation system (10), and 82.8% for allele-specific PCR assay (19). If our specimens and strains were subjected to analysis by the line probe assay (INNO-LiPA Rif.TB; Immogenetics, Ghent, Belgium), it is likely that all strains with RRDR mutations can be detected by negative signals for one or more of the wild-type probes S1 to S5. However, the line probe assay is inferior to sequencing in that the ability to assign the specific resistance mutation is limited. Thus, of the nine unique resistance mutations identified, the line probe assay would only be able to assign Ser531Leu, His526Asp, and Asp516Val. The confidence to predict RIF susceptibility from a no-mutation result, however, is likely to differ depending on the populations and might vary as new resistance mechanism emerges. In the present sample drawn from Southern China, all of the negative (i.e., no mutation) results represented “true susceptibility” since all of the RIFresistant isolates had at least one mutation in the RRDR, as previously described for isolates from Italy (22). In contrast, others have reported that some RIF-resistant strains had no mutations in the RRDR core: 2.6 to 10% in China (10, 35), 3.3 to 11% in other Asian countries (12, 23, 34), 3% in Australia (37), and 5% in Germany (11). In the present investigation, the most common substitutions among the RIF-resistant strains occurred at codons Ser-531 (59%) and His-526 (23.1%). This finding is similar to results compiled for 478 RIF-resistant M. tuberculosis isolates from various parts of the world, in which the frequencies of substitution at codons 531 and 526 were 41 and 36%, respectively (24). A model for implementation of a suggested strategy of directly applying the rpoB PCR-sequencing on respiratory specimens or on culture isolates is depicted in Fig. 2. Due to the significant costs of PCR and sequencing, respiratory specimens should be prioritized rather than tested universally. High priority should be given to circumstances in which the risk of MDR-TB is high, such as recurrent tuberculosis (18), previously treated cases (8), treatment default (14), irregular prior treatment (15), and smear positivity after 2 months of therapy (4). At this time, the service should only be offered selectively at the reference laboratory level. As a cost reduction measure, our experience demonstrates that an adequate result could be obtained by using one-quarter reaction volumes for DNA sequencing. However, a potential caveat is that in M. tuberculosis, resistance to RIF can occur at low frequency with only a few percent of the population being resistant. In the agar proportion method, a strain is defined as resistant if there is any growth of ⬎1% compared to no drug treatment. Thus, the cell population in a resistant strain is mixed with a small resistant subpopulation and a larger sensitive subpopulation. Depending on the population fractions sampled, sequencing-based or other molecule-based testing might yield a false-sensitive result if the resistant subpopulations are not included. Another limitation of the present method is that direct determination of RIF resistance was not possible for 40% of the 130 smear-

4442

YAM ET AL.

J. CLIN. MICROBIOL.

3. 4.

5. 6.

7. 8.

9.

10.

11.

12.

13.

14.

FIG. 2. Model for the implementation of genotypic detection of RIF resistance by rpoB PCR-sequencing. MTB, M. tuberculosis.

negative–culture-positive specimens. In a retrospective review of 29,057 tuberculosis cases, including 603 MDR-TB cases, it was found that 40% of patients with MDR-TB had negative smears in respiratory specimens at presentation (13). Hence, further studies to enhance the diagnostic performance of rpoB PCR-sequencing for this group of specimens are clearly warranted. ACKNOWLEDGMENTS This study was supported by a grant from the UDF Project-Research Centre of Emerging Infectious Diseases. We thank the staff at the clinical laboratories in the Queen Mary Hospital for technical support; P. K. Ip, Public Health Laboratory, Macau, Peoples Republic of China; and Y. K. Tam, Medical Laboratory, Chest and Lung Hospital, Guangzhou, Peoples Republic of China, for the provision of clinical isolates.

15.

16.

17.

18. 19.

20.

21.

REFERENCES 1. Barlett, J. G., K. J. Ryan, T. F. Smith, and W. R. Wilson. 1987. Cumitech 7A: laboratory diagnosis of lower respiratory tract infections. American Society for Microbiology, Washington, D.C. 2. Bogard, M., J. Vincelette, R. Antinozzi, R. Alonso, T. Fenner, J. Schirm, D. Aubert, C. Gaudreau, E. Sala, M. J. Ruiz-Serrano, H. Petersen, L. A. Oostendorp, and H. Burkardt. 2001. Multicenter study of a commercial, auto-

22.

23.

mated polymerase chain reaction system for the rapid detection of Mycobacterium tuberculosis in respiratory specimens in routine clinical practice. Eur. J. Clin. Microbiol. Infect. Dis. 20:724–731. Chan, C. M., K. Y. Yuen, K. S. Chan, W. C. Yam, K. H. Yim, W. F. Ng, and M. H. Ng. 1996. Single-tube nested PCR in the diagnosis of tuberculosis. J. Clin. Pathol. 49:290–294. Chavez Pachas, A. M., R. Blank, M. C. Smith Fawzi, J. Bayona, M. C. Becerra, and C. D. Mitnick. 2004. Identifying early treatment failure on category I therapy for pulmonary tuberculosis in Lima Ciudad, Peru. Int. J. Tuberc. Lung Dis. 8:52–58. Cheng, V. C., W. C. Yam, I. F. Hung, P. C. Woo, S. K. Lau, B. S. Tang, and K. Y. Yuen. 2004. Clinical evaluation of the polymerase chain reaction for the rapid diagnosis of tuberculosis. J. Clin. Pathol. 57:281–285. Drobniewski, F. A., S. A. Watterson, S. M. Wilson, and G. S. Harris. 2000. A clinical, microbiological and economic analysis of a national service for the rapid molecular diagnosis of tuberculosis and rifampicin resistance in Mycobacterium tuberculosis. J. Med. Microbiol. 49:271–278. Drobniewski, F. A., and S. M. Wilson. 1998. The rapid diagnosis of isoniazid and rifampicin resistance in Mycobacterium tuberculosis: a molecular story. J. Med. Microbiol. 47:189–196. Espinal, M. A., K. Laserson, M. Camacho, Z. Fusheng, S. J. Kim, R. E. Tlali, I. Smith, P. Suarez, M. L. Antunes, A. G. George, N. Martin-Casabona, P. Simelane, K. Weyer, N. Binkin, and M. C. Raviglione. 2001. Determinants of drug-resistant tuberculosis: analysis of 11 countries. Int. J. Tuberc. Lung Dis. 5:887–893. Espinal, M. A., A. Laszlo, L. Simonsen, F. Boulahbal, S. J. Kim, A. Reniero, S. Hoffner, H. L. Rieder, N. Binkin, C. Dye, R. Williams, M. C. Raviglione, et al. 2001. Global trends in resistance to antituberculosis drugs. N. Engl. J. Med. 344:1294–1303. Fan, X. Y., Z. Y. Hu, F. H. Xu, Z. Q. Yan, S. Q. Guo, and Z. M. Li. 2003. Rapid detection of rpoB gene mutations in rifampin-resistant Mycobacterium tuberculosis isolates in Shanghai by using the amplification refractory mutation system. J. Clin. Microbiol. 41:993–997. Heep, M., B. Brandstatter, U. Rieger, N. Lehn, E. Richter, S. Rusch-Gerdes, and S. Niemann. 2001. Frequency of rpoB mutations inside and outside the cluster I region in rifampin-resistant clinical Mycobacterium tuberculosis isolates. J. Clin. Microbiol. 39:107–110. Hirano, K., C. Abe, and M. Takahashi. 1999. Mutations in the rpoB gene of rifampin-resistant Mycobacterium tuberculosis strains isolated mostly in Asian countries and their rapid detection by line probe assay. J. Clin. Microbiol. 37:2663–2666. Kam, K. M., C. W. Yip, L. W. Tse, O. C. Leung, L. P. Sin, M. Y. Chan, and W. S. Wong. 2002. Trends in multidrug-resistant Mycobacterium tuberculosis in relation to sputum smear positivity in Hong Kong, 1989–1999. Clin. Infect. Dis. 34:324–329. Kimerling, M. E., A. Slavuckij, S. Chavers, G. G. Peremtin, T. Tonkel, O. Sirotkina, V. Golubchikova, and A. Baddeley. 2003. The risk of MDR-TB and polyresistant tuberculosis among the civilian population of Tomsk City, Siberia, 1999. Int. J. Tuberc. Lung Dis. 7:866–872. Kritski, A. L., L. S. Rodrigues de Jesus, M. K. Andrade, E. WerneckBarroso, M. A. Vieira, A. Haffner, and L. W. Riley. 1997. Retreatment tuberculosis cases. Factors associated with drug resistance and adverse outcomes. Chest 111:1162–1167. Leung, E. T., K. M. Kam, A. Chiu, P. L. Ho, W. H. Seto, K. Y. Yuen, and W. C. Yam. 2003. Detection of katG Ser315Thr substitution in respiratory specimens from patients with isoniazid-resistant Mycobacterium tuberculosis using PCR-RFLP. J. Med. Microbiol. 52:999–1003. Levidiotou, S., G. Vrioni, E. Galanakis, E. Gesouli, C. Pappa, and D. Stefanou. 2003. Four-year experience of use of the Cobas Amplicor system for rapid detection of Mycobacterium tuberculosis complex in respiratory and nonrespiratory specimens in Greece. Eur. J. Clin. Microbiol. Infect. Dis. 22:349–356. Liu, Z., K. L. Shilkret, and L. Finelli. 1998. Epidemiology of drug-resistant tuberculosis in New Jersey from 1991 to 1995. Int. J. Epidemiol. 27:121–126. Mokrousov, I., T. Otten, B. Vyshnevskiy, and O. Narvskaya. 2003. Allelespecific rpoB PCR assays for detection of rifampin-resistant Mycobacterium tuberculosis in sputum smears. Antimicrob. Agents Chemother. 47:2231– 2235. National Committee for Clinical Laboratory Standards. 2000. Susceptibility testing of Mycobacteria, Nocardia and other Aerobic Actinomycetes. Tentative standard MT24-T2. National Committee for Clinical Laboratory Standards, Wayne, Pa. Nolte, F. S., and B. Metchock. 1995. Mycobacterium, p. 400–437. In P. R. Murray, E. J. Baron, M. A. Pfaller, F. C. Tenover, and R. H. Yolken (ed.), Manual of clinical microbiology. American Society for Microbiology, Washington, D.C. Pozzi, G., M. Meloni, E. Iona, G. Orru, O. F. Thoresen, M. L. Ricci, M. R. Oggioni, L. Fattorini, and G. Orefici. 1999. rpoB mutations in multidrugresistant strains of Mycobacterium tuberculosis isolated in Italy. J. Clin. Microbiol. 37:1197–1199. Qian, L., C. Abe, T. P. Lin, M. C. Yu, S. N. Cho, S. Wang, and J. T. Douglas.

VOL. 42, 2004

24. 25.

26.

27. 28.

29. 30. 31.


2002. rpoB genotypes of Mycobacterium tuberculosis Beijing family isolates from East Asian countries. J. Clin. Microbiol. 40:1091–1094. Ramaswamy, S., and J. M. Musser. 1998. Molecular genetic basis of antimicrobial agent resistance in Mycobacterium tuberculosis: 1998 update. Tuberc. Lung Dis. 79:3–29. Ruiz, M., M. J. Torres, A. C. Llanos, A. Arroyo, J. C. Palomares, and J. Aznar. 2004. Direct detection of rifampin- and isoniazid-resistant Mycobacterium tuberculosis in auramine-rhodamine-positive sputum specimens by real-time PCR. J. Clin. Microbiol. 42:1585–1589. Schilke, K., K. Weyer, G. Bretzel, B. Amthor, J. Brandt, V. Sticht-Groh, P. B. Fourie, and W. H. Haas. 1999. Universal pattern of rpoB gene mutations among multidrug-resistant isolates of Mycobacterium tuberculosis complex from Africa. Int. J. Tuberc. Lung Dis. 3:620–626. Sintchenko, V., W. K. Chew, P. J. Jelfs, and G. L. Gilbert. 1999. Mutations in rpoB gene and rifabutin susceptibility of multidrug-resistant Mycobacterium tuberculosis strains isolated in Australia. Pathology 31:257–260. Telenti, A., N. Honore, C. Bernasconi, J. March, A. Ortega, B. Heym, H. E. Takiff, and S. T. Cole. 1997. Genotypic assessment of isoniazid and rifampin resistance in Mycobacterium tuberculosis: a blind study at reference laboratory level. J. Clin. Microbiol. 35:719–723. Telenti, A., P. Imboden, F. Marchesi, D. Lowrie, S. Cole, M. J. Colston, L. Matter, K. Schopfer, and T. Bodmer. 1993. Detection of rifampicin-resistance mutations in Mycobacterium tuberculosis. Lancet 341:647–650. Valim, A. R., M. L. Rossetti, M. O. Ribeiro, and A. Zaha. 2000. Mutations in the rpoB gene of multidrug-resistant Mycobacterium tuberculosis isolates from Brazil. J. Clin. Microbiol. 38:3119–3122. Williams, D. L., L. Spring, T. P. Gillis, M. Salfinger, and D. H. Persing. 1998.

32.

33.

34.

35. 36.

37.

4443

Evaluation of a polymerase chain reaction-based universal heteroduplex generator assay for direct detection of rifampin susceptibility of Mycobacterium tuberculosis from sputum specimens. Clin. Infect. Dis. 26:446–450. Yam, W. C., V. C. Cheng, W. T. Hui, L. N. Wang, W. H. Seto, and K. Y. Yuen. 2004. Direct detection of Mycobacterium tuberculosis in clinical specimens using single-tube biotinylated nested polymerase chain reaction-enzyme linked immunoassay (PCR-ELISA). Diagn. Microbiol. Infect. Dis. 48:271– 275. Yam, W. C., K. Y. Yuen, and W. H. Seto. 1998. Direct detection of Mycobacterium tuberculosis in respiratory specimens using an automated DNA amplification assay and a single tube nested polymerase chain reaction (PCR). Clin. Chem. Lab. Med. 36:597–599. Yang, B., H. Koga, H. Ohno, K. Ogawa, M. Fukuda, Y. Hirakata, S. Maesaki, K. Tomono, T. Tashiro, and S. Kohno. 1998. Relationship between antimycobacterial activities of rifampicin, rifabutin, and KRM-1648 and rpoB mutations of Mycobacterium tuberculosis. J. Antimicrob. Chemother. 42:621– 628. Yue, J., W. Shi, J. Xie, Y. Li, E. Zeng, and H. Wang. 2003. Mutations in the rpoB gene of multidrug-resistant Mycobacterium tuberculosis isolates from China. J. Clin. Microbiol. 41:2209–2212. Yuen, K. Y., W. C. Yam, L. P. Wong, and W. H. Seto. 1997. Comparison of two automated DNA amplification systems with a manual one-tube nested PCR assay for diagnosis of pulmonary tuberculosis. J. Clin. Microbiol. 35: 1385–1389. Yuen, L. K., D. Leslie, and P. J. Coloe. 1999. Bacteriological and molecular analysis of rifampin-resistant Mycobacterium tuberculosis strains isolated in Australia. J. Clin. Microbiol. 37:3844–3850.