See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/256452078

Prospective multicentre evaluation of the direct nitrate reductase assay for

the rapid detection of extensively drug-resistant tuberculosis

Article in Journal of Antimicrobial Chemotherapy · September 2013

DOI: 10.1093/jac/dkt353 · Source: PubMed

CITATIONS READS

20 121

16 authors, including:

Anandi Martin Belen ROCIO Imperiale

Université Catholique de Louvain - UCLouvain National Academy of Medicine, IMEX-CONICET
166 PUBLICATIONS 5,596 CITATIONS 43 PUBLICATIONS 615 CITATIONS

SEE PROFILE SEE PROFILE

Ahmet Yilmaz Coban Aamer Ikram

Ondokuz Mayıs Üniversitesi Armed Forces Institute of Pathology
109 PUBLICATIONS 1,145 CITATIONS 395 PUBLICATIONS 3,497 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Aamer Ikram on 30 December 2015.

The user has requested enhancement of the downloaded file.

J Antimicrob Chemother 2014; 69: 441 – 444
doi:10.1093/jac/dkt353 Advance Access publication 5 September 2013

Prospective multicentre evaluation of the direct nitrate reductase assay

for the rapid detection of extensively drug-resistant tuberculosis
Anandi Martin1*, Belen Imperiale2, Pascaline Ravolonandriana3, Ahmet Yilmaz Coban4, Alper Akgunes5,
Aamer Ikram6, Luqman Satti6, Mathieu Odoun7, Pooja Pandey8, Manvi Mishra8, Dissou Affolabi7, Urvashi Singh8,
Voahangy Rasolofo3, Nora Morcillo2, Peter Vandamme1 and Juan Carlos Palomino1

1
Laboratory of Microbiology, Department of Biochemistry and Microbiology, Ghent University, Ghent, Belgium; 2Laboratorio de Referencia del
Programa de Control de Tuberculosis, Hospital Dr Antonio Cetrángolo, Buenos Aires, Argentina; 3Institut Pasteur de Madagascar,
Antananarivo, Madagascar; 4Department of Medical Microbiology, Medical School, Ondokuz Mayis University, Samsun, Turkey; 5Clinical
Microbiology Laboratory, Chest Diseases Hospital, Samsun, Turkey; 6Department of Pathology, Combined Military Hospital, Dera Ismail Khan,
Pakistan; 7Laboratoire de Référence des Mycobactéries, Cotonou, Benin; 8Department of Microbiology, All India Institute of Medical Sciences,

Downloaded from http://jac.oxfordjournals.org/ by guest on December 22, 2015

New Delhi, India

*Corresponding author. Tel: +32-9-264-51-20; Fax: +32-9-264-50-92; E-mail: [email protected]

Received 16 June 2013; returned 23 July 2013; revised 30 July 2013; accepted 5 August 2013

Objectives: To perform a multicentre study evaluating the performance of the direct nitrate reductase assay (NRA)
for the detection of multidrug-resistant (MDR) and extensively drug-resistant (XDR) tuberculosis in sputum samples.
Methods: The study was conducted in six laboratories performing tuberculosis diagnosis that were located in six dif-
ferent countries. The NRA was performed directly on sputum samples in parallel with the reference method used at
each site. Detection of resistance was performed for rifampicin, isoniazid, ofloxacin and kanamycin.
Results: Excellent agreement was obtained for all drugs tested at the majority of sites. The accuracy was 93.7%–
100% for rifampicin, 88.2%–100% for isoniazid, 94.6%–100% for ofloxacin and 100% for kanamycin. The majority
of NRA results were available at day 21 for sites 1, 2 and 5. Site 3 had a turnaround time of 13.9 days, at site 4 it was
18.4 days and at site 6 it was 16.2 days. The contamination rate ranged between 2.5% and 12%.
Conclusions: Rapid detection of drug resistance by the direct NRA on sputum smear-positive samples was accurate
and easy to implement in clinical diagnostic laboratories, making it a good alternative for rapid screening for MDR and
XDR tuberculosis.

Keywords: multidrug resistance, nitratase, Mycobacterium tuberculosis

Introduction still too expensive or have logistical problems preventing their

broad adoption in low-income countries with a significant number
The increasing incidence of multidrug-resistant (MDR) and exten- of drug-resistant cases.
sively drug-resistant (XDR) tuberculosis (TB) highlights the need The nitrate reductase assay (NRA) is a rapid, low-cost, pheno-
for rapid, affordable and accurate methods for testing susceptibil- typic method performed on solid media and based on the metabol-
ity to first- and second-line anti-TB drugs. XDR Mycobacterium tu- ic activity of M. tuberculosis;8 it has been endorsed by the WHO for
berculosis strains, by being resistant to several second-line drugs the rapid detection of MDR-TB.9 The aim of this study was to evalu-
in addition to rifampicin and isoniazid, are a new threat to the ef- ate in a multicentre study the feasibility of using the NRA for the
fective control of the disease.1 The current methods for drug sus- simultaneous detection of MDR- and XDR-TB through the detection
ceptibility testing of M. tuberculosis are either slow or costly. of resistance to isoniazid, rifampicin, ofloxacin and kanamycin dir-
Löwenstein –Jensen (LJ) medium is still the most commonly ectly from sputum samples.
used and inexpensive solid medium for drug resistance detection
in many clinical laboratories. Unfortunately, conventional drug Methods
susceptibility testing methods based on LJ medium are time con-
suming, taking up to 6 weeks to give results.2,3 This is mainly due Study design and setting
to the slow growth of M. tuberculosis. In order to reduce this turn- This prospective study was carried out in six laboratories situated in six differ-
around time, commercial broth-based systems and molecular ent countries: Department of Microbiology, All India Institute of Medical
tests have been developed.4 – 7 Nevertheless, these systems are Sciences, New Delhi, India; Department of Pathology, Combined Military

# The Author 2013. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: [email protected]

441
Martin et al.

Hospital, Dera Ismail Khan, Pakistan; Laboratorio de Referencia del Programa NRA
de Control de Tuberculosis, Hospital Dr Antonio Cetrángolo, Buenos Aires, Ar-
The LJ–NRA medium was prepared in-house and the method was carried out
gentina; Clinical Microbiology Laboratory, Chest Diseases Hospital, Samsun,
as previouslydescribed11 with a minor modification in reading timepoints. LJ–
Turkey; Institut Pasteur de Madagascar, Antananarivo, Madagascar; and
NRA tubes contained 1 g/L potassium nitrate (KNO3) and critical concentra-
Laboratoire de Référence des Mycobactéries, Cotonou, Benin. The participat-
tions of 40.0 mg/L rifampicin, 0.2 mg/L isoniazid, 2.0 mg/L ofloxacin and
ing laboratories were randomly numbered from 1 to 6 for the purpose of this
30 mg/L kanamycin. An additional LJ–NRA tube without drugs contained
publication. The inclusion criteria were cough for .3 weeks, age .12 years,
0.5 g/L para-aminobenzoic acid (PNB) to identify the M. tuberculosis complex
suspected pulmonary TB, and suspected treatment failure, suspected
and three LJ–NRA drug-free tubes served as a growth control. Drug- and
relapse or treatment default. Only those samples that were positive by
PNB-containing tubes were inoculated with 0.2 mL of the concentrated and
Ziehl–Neelsen smear microscopy were included in the study. A total of 265
decontaminated sputum, while growth controls were inoculated with
smear-positive (≥1+) sputum samples were collected prospectively at the
0.2 mL of a 1:10 dilution of the same sample. After 10 days of incubation
six sites from January 2012 to October 2012. All participating laboratories
at 378C, the NRA was developed by adding 0.5 mL of freshly prepared Griess
had experience with the NRA or were previously trained.
reagent consisting of 1 part 50% (v/v) hydrochloric acid, 2 parts 0.2% (w/v)
sulfanilamide and 2 parts 0.1% (w/v) N-(1-naphthyl)ethylenediamine dihy-
Sputum samples drochloride to one of the growth controls. If any pink colour appeared, the
Griess reagent was added to all drug-containing tubes. If there was no
Smear microscopy was performed directly on all sputum samples. Those colour change, the remaining tubes were reincubated and the procedure
that were smear positive were selected and decontaminated by the repeated at days 14 and 28. An isolate was considered resistant if any

Downloaded from http://jac.oxfordjournals.org/ by guest on December 22, 2015

sodium hydroxide –N-acetyl-L-cysteine method.2 Samples were then con- colour change was observed in the drug-containing tube. If the control was
centrated by centrifugation at 3200 g for 20 min. The supernatant was dis- negative at day 28, the NRA was considered invalid. All tests were performed
carded, and the remaining sediment was resuspended in 3 mL of sterile blinded. Identification of the M. tuberculosis complex was made by reading
distilled water and used to inoculate, in parallel, either LJ medium or the the PNB-containing tube. If no colour change occurred, it was considered to
Bactec MGIT 960 Mycobacterial Detection System (as the reference be M. tuberculosis complex.
method) for indirect drug susceptibility testing, and NRA tubes containing
antibiotics for direct testing.
Statistical analysis
LJ proportion method MedCalc statistical software version 12.3.0.0 (Mariakerke, Belgium) was
used to calculate the statistical measures of sensitivity, specificity, accur-
The proportion method was performed on LJ medium according to the
acy, positive predictive value (PPV) and negative predictive value (NPV) of
standard procedure, with the recommended critical concentrations of
the direct NRA using the LJ proportion method or the MGIT 960 system as
40 mg/L rifampicin, 0.2 mg/L isoniazid, 2 mg/L ofloxacin and 30 mg/L
the reference test. The time to positivity (TTP) of the NRA was also recorded.
kanamycin. LJ tubes were incubated at 378C and read at 28 and 42 days.3

MGIT 960 system

Results
Testing using the MGIT 960 system was performed according to the manu- For all drugs, excellent agreement was obtained between the direct
facturer’s instructions, with 1 mg/L rifampicin and 0.1 mg/L isoniazid. For NRA and the gold standard method used at each site – either the LJ
the second-line drugs, 2 mg/L ofloxacin and 2.5 mg/L kanamycin were proportion method or the MGIT 960 system. The sole exception was
tested according to Martin et al.10 ofloxacin at one site. Table 1 shows for each site the number of

Table 1. Susceptibility results for the direct NRA compared with the MGIT 960 system or LJ proportion method for each site

MGIT 960 system or LJ proportion method

rifampicin isoniazid ofloxacin kanamycin

Site Direct NRA R S R S R S R S

1 R 18 3 20 1 11 0 7 0
S 0 31 0 31 0 41 0 45
2 R 12 0 15 0 1 0 1 0
S 1 37 2 33 0 49 0 49
3 R 4 0 9 0 2 0 5 0
S 0 53 0 48 0 55 0 52
4 R 0 0 6 0 2 1 0 0
S 0 52 2 44 1 48 0 52
5 R 3 0 4 1 0 0 0 0
S 1 12 1 11 0 17 0 17
6 R 2 0 2 0 0 2 0 0
S 0 35 0 35 0 35 0 37

R, resistant; S, susceptible.

442
Direct nitrate reductase assay for XDR-TB JAC
Table 2. Sensitivity, specificity, PPV, NPV and accuracy of the direct NRA for each drug tested at each study site

Rifampicin (%) Isoniazid (%) Ofloxacin (%) Kanamycin (%)

Site sens spec PPV NPV A sens spec PPV NPV A sens spec PPV NPV A sens spec PPV NPV A

1 100 91.1 85.7 100 94.2 100 96.8 95.2 100 98.0 100 100 100 100 100 100 100 100 100 100
2 92.3 100 100 97.3 98.0 88.2 100 100 94.2 96.0 100 100 100 100 100 100 100 100 100 100
3 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100
4 NA 100 NA NA 100 75 100 100 95.6 96.2 66.7 98 66.6 97.9 96.2 NA 100 NA 100 100
5 75 100 100 92.3 93.7 80 91.6 80 91.6 88.2 NA 100 NA 100 100 NA 100 NA 100 100
6 100 100 100 100 100 100 100 100 100 100 NA 94.6 NA 100 94.6 NA 100 NA 100 100

sens, sensitivity; spec, specificity; A, accuracy; NA, not applicable.

resistant and susceptible samples detected by the NRA compared Middlebrook agar by the proportion method is very slow, requiring

Downloaded from http://jac.oxfordjournals.org/ by guest on December 22, 2015

with results obtained by the reference method used at the site. 3 – 6 weeks to produce results. The development of affordable
For rifampicin, only site 1 had false resistant results and two and rapid drug susceptibility testing methods for low-resource
sites (site 2 and site 5) had one false susceptible result each. Site countries is then a continuing priority. The NRA described here
5 had a sample contaminated for rifampicin and was discarded has the advantage of a shorter TTP (10 – 14 days) compared
for the analysis. For isoniazid, two sites (site 1 and 5) had one with the conventional proportion method. The NRA is based on
false resistant result each, while sites 2 and 4 had two false suscep- the detection of nitrate reduction as an indicator of growth, and
tible and site 5 one false susceptible result. the results are therefore obtained before any macroscopic
For ofloxacin, site 4 had one false resistant and one false suscep- growth can be visually detected.
tible result. Because of the very low number of strains resistant A number of studies have reported the usefulness of the NRA
to ofloxacin by the NRA, the sensitivity was lower for this drug. for determining susceptibility or resistance to rifampicin and iso-
Site 6 found two samples resistant to ofloxacin by the NRA; niazid, the two most important drugs for the treatment of TB, and
however, the proportion method found those samples to be sus- has shown high sensitivity and specificity.12 – 14 The NRA also has
ceptible. For kanamycin, all results were concordant at all sites. the potential to be used for the detection of resistance to second-
Site 4 did not find any samples resistant to either rifampicin or line drugs. In 2005, our group reported the first evaluation of the
kanamycin during the study period. Similarly, site 5 did not have NRA for the detection of ofloxacin resistance and found complete
any ofloxacin- or kanamycin-resistant samples. concordance with the proportion method.15 Rosales et al.16 eval-
Based on these results, Table 2 shows the specificity, sensitivity, uated the NRA for the rapid detection of resistance to ofloxacin
PPV, NPV and accuracy of the NRA method for all sites. For rifampi- and kanamycin in Honduras. They found good specificity for
cin the accuracy ranged between 93.7% and 100%, for isoniazid both drugs, but lower sensitivity for detecting resistance to kana-
the accuracy ranged between 88.2% and 100%, for ofloxacin the mycin. They also showed how the NRA could easily be adapted to
accuracy ranged between 94.6% and 100%, and for kanamycin the local conditions of a diagnostic laboratory that was already
the accuracy was 100%. The majority of NRA results were available using the proportion method on LJ medium. More recently, Visa-
at day 21 for sites 1, 2 and 5, with an average of 72% positivity at lakshi et al.17 evaluated the NRA for rapid detection of resistance
day 21, 12% at day 14 and 16% at day 28. Site 3 had a TTP of to kanamycin, ethambutol, ofloxacin, cycloserine and para-
13.9 days, at site 6 it was 16.2 days and at site 4 it was aminosalicylic acid. They obtained sensitivity that ranged
18.4 days. For site 3, out of 61 consecutive samples, 2 did not between 86.4% and 100%, and specificity ranging between
yield growth and another 2 were identified as non-tuberculous 98.4% and 100%.
mycobacteria (Mycobacterium avium). The contamination rate In the present study, the sensitivity and specificity for detect-
was between 2.5% and 12% for the NRA. ing resistance to rifampicin and isoniazid were excellent and
.94% for the second-line drugs. This is the first multicentre
study that confirms that the direct NRA can also be used to
Discussion screen XDR-TB. This study was limited by the relatively small
We performed the first multicentre prospective evaluation to assess number of XDR-TB cases found. The NRA has been adapted for
the performance of the direct NRA assay for the detection of MDR the direct detection of MDR-TB using smear-positive sputum
and XDR-TB. The assay was easy to implement in the field setting, samples in a number of countries such as Peru, Argentina,
showed low or acceptable contamination levels and a shorter turn- Benin, Brazil, India, Nigeria and Pakistan.16,18 – 23 The main advan-
around time than the proportion method on LJ medium, and a tage of the direct NRA is its much shorter turnaround time. Also,
similar TTP compared with the MGIT 960 system. The direct NRA the simplicity and cost-effectiveness of the test is suitable for
showed overall good sensitivity and specificity for the four drugs large-scale surveillance studies in resource-limited settings.
tested. Furthermore, incorporating PNB in the assay allowed very It furthermore offers the advantage of using solid media,
simple identification of the M. tuberculosis complex. which minimizes cross-contamination and biohazard risks.
Although considered as the gold standard, it is well known that Additional studies of direct NRA testing for second-line drugs
drug susceptibility testing performed on LJ medium or are warranted.

443
Martin et al.

Our study was limited by the low number of drug-resistant 9 WHO. Strategic and Technical Advisory Group for Tuberculosis (STAG-TB).
samples found in the different countries. Nevertheless, the direct Report of the Ninth Meeting. http://www.who.int/tb/advisory_bodies/
NRA can be used as a rapid and low-cost screening method to stag_tb_report_2009.pdf (4 June 2013, date last accessed).
detect drug resistance. 10 Martin A, von Groll A, Fissette K et al. Rapid detection of Mycobacterium
In conclusion, rapid detection of drug resistance by the direct tuberculosis resistance to second-line drugs by use of the manual
NRA on sputum smear-positive samples was accurate and easy mycobacterium growth indicator tube system. J Clin Microbiol 2008; 12:
to implement in clinical diagnostic laboratories, making it a good 3952– 6.
option for rapid screening for MDR- and XDR-TB. 11 Martin A, Montoro E, Lemus D et al. Multicenter evaluation of the nitrate
reductase assay for drug resistance detection of Mycobacterium
tuberculosis. J Microbiol Methods 2005; 63: 145–50.
12 Lemus D, Montoro E, Echemendı́a M et al. Nitrate reductase assay for
Acknowledgements detection of drug resistance in Mycobacterium tuberculosis: simple and
We thank the Damien Foundation, Belgium, for their support. inexpensive method for low-resource laboratories. J Med Microbiol 2006;
55: 861– 3.
13 Martin A, Portaels F, Palomino JC. Colorimetric redox-indicator methods for
Funding the rapid detection of multidrug resistance in Mycobacterium tuberculosis: a
The study was partially funded by the Damien Foundation, Belgium. B. I. has systematic review and meta-analysis. J Antimicrob Chemother 2007; 59:

Downloaded from http://jac.oxfordjournals.org/ by guest on December 22, 2015

a fellowship of the ‘Consejo Nacional de Investigaciones Cientı́ficas y Técni- 175–83.
cas’ (CONICET), Argentina. 14 Shin SS, Asencios L, Yagui M et al. Impact of rapid drug susceptibility
testing for tuberculosis: program experience in Lima, Peru. Int J Tuberc
Lung Dis 2012; 16: 1538 –43.
Transparency declarations 15 Martin A, Palomino JC, Portaels F. Rapid detection of ofloxacin
None to declare. resistance in Mycobacterium tuberculosis by two low-cost colorimetric
methods: resazurin and nitrate reductase assays. J Clin Microbiol 2005;
43: 1612 –6.
References 16 Rosales S, Pineda-Garcı́a L, Andino N et al. Evaluation of the nitrate
1 Klopper M, Warren RM, Hayes Cet al. Emergence and spread of extensively reductase assay for rapid detection of extensively drug-resistant tuberculosis.
and totally drug-resistant tuberculosis, South Africa. Emerg Infect Dis 2013; Int J Tuberc Lung Dis 2009; 13: 1542–9.
19: 449– 55. 17 Visalakshi P, Arora J, Behera D. Evaluation of nitrate reductase assay for
2 Kent PT, Kubica GP. Public Health Mycobacteriology. A Guide for the Level rapid detection of resistance to second-line drugs in Mycobacterium
III Laboratory. Atlanta, GA: CDC, 1985. tuberculosis in a tertiary care hospital. Diagn Microbiol Infect Dis 2010; 68:
3 Canetti G, Fox W, Khomenko A et al. Advances in techniques of testing 236–40.
mycobacterial drug sensitivity, and the use of sensitivity tests in 18 Satti L, Ikram A, Palomino JCet al. Field evaluation of the direct detection
tuberculosis control programmes. Bull World Health Organ 1969; 41: 21 –43. of multidrug resistant Mycobacterium tuberculosis by nitrate reductase
4 Bemer P, Palicova F, Rüsch-Gerdes S et al. Multicenter evaluation of fully assay on 7H11 agar. Tuberculosis (Edinb) 2013; 93: 308–11.
automated BACTEC Mycobacteria Growth Indicator Tube 960 system for 19 Solis LA, Shin SS, Han LL et al. Validation of a rapid method for detection
susceptibility testing of Mycobacterium tuberculosis. J Clin Microbiol 2002; of M. tuberculosis resistance to isoniazid and rifampin in Lima, Peru. Int J
40: 150– 4. Tuberc Lung Dis 2005; 9: 760–4.
5 Angeby KA, Werngren J, Toro JC et al. Evaluation of the BacT/ALERT 3D 20 Musa HR, Ambroggi M, Souto A et al. Drug susceptibility testing of
system for recovery and drug susceptibility testing of Mycobacterium Mycobacterium tuberculosis by a nitrate reductase assay applied directly
tuberculosis. Clin Microbiol Infect 2003; 9: 1148 –52. on microscopy-positive sputum samples. J Clin Microbiol 2005; 43:
6 Rossau R, Traore H, De Beenhouwer H et al. Evaluation of the INNO-LiPA 3159– 61.
Rif. TB assay, a reverse hybridization assay for the simultaneous detection 21 Mishra B, Muralidharan S, Srinivasa H. Direct drug susceptibility testing of
of Mycobacterium tuberculosis complex and its resistance to rifampin. Mycobacterium tuberculosis to primary anti-tubercular drugs by nitrate
Antimicrob Agents Chemother 1997; 41: 2093– 8. reductase assay. Indian J Pathol Microbiol 2009; 52: 343– 4.
7 Kontsevaya I, Mironova S, Nikolayevskyy V et al. Evaluation of two 22 Affolabi D, Odoun M, Martin A et al. Evaluation of direct detection of
molecular assays for rapid detection of Mycobacterium tuberculosis Mycobacterium tuberculosis rifampin resistance by a nitrate reductase
resistance to fluoroquinolones in high-tuberculosis and -multidrug- assay applied to sputum samples in Cotonou, Benin. J Clin Microbiol 2007;
resistance settings. J Clin Microbiol 2011; 49: 2832 –7. 45: 2123 –5.
8 Angeby KA, Klintz L, Hoffner SE. Rapid and inexpensive drug susceptibility 23 Shikama ML, Ferro e Silva R, Villela G et al. Multicentre study of nitrate
testing of Mycobacterium tuberculosis with a nitrate reductase assay. J Clin reductase assay for rapid detection of rifampicin-resistant M. tuberculosis.
Microbiol 2002; 40: 553–5. Int J Tuberc Lung Dis 2009; 13: 377– 80.

444

View publication stats