Comparative Evaluation of GenoType MTBDRplus Line Probe Assay with Solid Culture Method in Early Diagnosis of Multidrug Resistant Tuberculosis (MDR-TB) at a Tertiary Care Centre in India

Background The objectives of the study were to compare the performance of line probe assay (GenoType MTBDRplus) with solid culture method for an early diagnosis of multidrug resistant tuberculosis (MDR-TB), and to study the mutation patterns associated with rpoB, katG and inhA genes at a tertiary care centre in north India. Methods In this cross-sectional study, 269 previously treated sputum-smear acid-fast bacilli (AFB) positive MDR-TB suspects were enrolled from January to September 2012 at the All India Institute of Medical Sciences hospital, New Delhi. Line probe assay (LPA) was performed directly on the sputum specimens and the results were compared with that of conventional drug susceptibility testing (DST) on solid media [Lowenstein Jensen (LJ) method]. Results DST results by LPA and LJ methods were compared in 242 MDR-TB suspects. The LPA detected rifampicin (RIF) resistance in 70 of 71 cases, isoniazid (INH) resistance in 86 of 93 cases, and MDR-TB in 66 of 68 cases as compared to the conventional method. Overall (rifampicin, isoniazid and MDR-TB) concordance of the LPA with the conventional DST was 96%. Sensitivity and specificity were 98% and 99% respectively for detection of RIF resistance; 92% and 99% respectively for detection of INH resistance; 97% and 100% respectively for detection of MDR-TB. Frequencies of katG gene, inhA gene and combined katG and inhA gene mutations conferring all INH resistance were 72/87 (83%), 10/87 (11%) and 5/87 (6%) respectively. The turnaround time of the LPA test was 48 hours. Conclusion The LPA test provides an early diagnosis of monoresistance to isoniazid and rifampicin and is highly sensitive and specific for an early diagnosis of MDR-TB. Based on these findings, it is concluded that the LPA test can be useful in early diagnosis of drug resistant TB in high TB burden countries.


Introduction
MDR-TB poses a great threat to the TB control programmes worldwide [1]. India accounts for more than 25% of the world's incident cases of tuberculosis [2]. There are an estimated 99,000 MDR-TB patients among incident total TB cases in India [3]. Early diagnosis of rifampicin (RIF) and isoniazid (INH) drug-resistant Mycobacterium tuberculosis (Mtb) is essential for efficient treatment and control of MDR-TB.
Solid and liquid culture methods for drug DST of Mtb are time consuming requiring weeks to months in providing the results. Further, contamination rates with conventional culture and DST are high. The World Health Organization (WHO), Geneva and Foundation for Innovative New Diagnostics (FIND), Geneva have endorsed the use of LPA test (GenoType MTBDRplus, Nehren, Germany) for rapid detection of MDR-TB directly from smearpositive sputum specimens and Mycobacterium tuberculosis cultures [4]. However, clinical utility of the test varies with the prevalence of particular mutations (incorporated in the test) in different geographical regions. There are limited data on the performance of this test from high TB burden countries including India. The LPA (MTBDRplus) detects mutations associated with the rpoB gene for RIF resistance, katG genes for high level INH resistance, and the inhA regulatory region gene for low-level INH resistance [5]. The aim of the study was to compare the performance of LPA test (GenoType MTBDRplus assay) with solid culture method among MDR-TB suspects for an early diagnosis of MDR-TB and monoresistance to isoniazid and rifampicin at a tertiary care centre in north India.

Study Subjects
The study was approved by the Institute Ethics Committee, All India Institute of Medical Sciences, New Delhi. Written informed consent was obtained from all patients. The patients were recruited from Chest Clinics of six districts in Delhi, medical outpatient department and DOT (Directly Observed Treatment) Centre of the All India Institute of Medical Sciences (AIIMS) hospital, New Delhi. All mycobacterial investigations were carried out at the Tuberculosis Laboratory of the Department of Internal Medicine. The laboratory is accredited for carrying out conventional DST and LPA tests by the National Mycobacteriology Accreditation System of Central TB Division, Ministry of Health and Family Welfare, Govt. of India. Sputum specimens from patients with a previous history of pulmonary TB treatment were subjected to microscopy. The smears were graded according to the number of bacilli seen on the slide, as per guidelines of the Revised National Tuberculosis Control Programme (RNTCP) of India [6]. All MDR-TB suspects from January to September 2012 with sputum bacillary load $1+ were enrolled in the study.

Specimen Collection and Processing
Two sputum samples (spot and morning) were collected from each patient in 50 ml wide-mouthed sterile falcon tubes according to the revised National Guidelines [7]. Both sputum samples were subjected to smear examination and culture. The LPA test was done in one of the samples having a higher bacillary load. The sputum specimens were handled in class II biosafety cabinet in a bio-safety level (BSL)-3 laboratory and sputum specimens were decontaminated by N-acetyl-L-cysteine and sodium hydroxide (NALC-NaOH) method [8]. Subsequently, the sediments were suspended in 1-1.5 ml sterile phosphate buffer (pH 6.8) and two bottles of Lowenstein-Jensen medium were inoculated with each sample. 500 ml of the processed sample was used for DNA isolation in a screw capped tube.

Conventional Drug Susceptibility Testing
The DST was carried out in Lowenstein-Jensen solid media by economic variant of 1% proportion method according to the standard operating procedure of RNTCP [9]. Rifampicin and isoniazid were tested with concentrations of 40 mg/ml and 0.2 mg/ ml respectively. All isolates were identified as Mycobacterium tuberculosis by their slow growth rate, colony morphology, inability to grow on L-J media containing p-nitrobenzoic acid (500 mg/ml), and niacin positive and catalase negative tests. Any strain with 1% (the critical proportion) of bacilli resistant to any of the two drugsrifampicin and isoniazid was classified as resistant to that drug.

Line Probe Assay
The GenoType MTBDRplus line probe assay was performed according to the manufacturer's (Hain Lifescience, Nehren, Germany) instructions. Three steps for LPA test included, DNA extraction, multiplex polymerase chain reaction (PCR) amplification and reverse hybridization. These steps were carried out in three separate rooms with restricted access and unidirectional workflow [10]. Mycobacterial DNA was extracted in BSL-3 laboratory according to manufacturer's instructions. Briefly, 500 ml of decontaminated sputum sample was centrifuged at 10,000Xg for 15 min, the supernatant was discarded and the pellet was resuspended in 100 ml sterile distilled water. The specimen was then heat killed at 95uC for 20 min in water bath. This was followed by sonication for 15 min and centrifugation at 13,000Xg for 8 min. Five ml of the DNA supernatant was used for PCR while the remainder was stored at 220uC. Master mixture for amplification consisted of 35 ml primer nucleotide mixture (provided with kit), 5 ml of 10XPCR buffer with 15 mM MgCl 2 , 2 ml of 25 mM MgCl 2 , 0.2 ml (1 U) of HotStarTaq DNA polymerase (Hain Lifescience, Nehren, Germany), 3 ml nuclease free molecular grade water and 5 ml of DNA supernatant in a final volume of 50 ml. The amplification protocol consisted of 15 min of denaturation at 95uC, followed by 10 cycles comprising denaturation at 95uC for 30 sec and 58uC for 2 min. This was followed by 20 cycles comprising 95uC for 25 sec, 53uC for 40 sec and 70uC for 40 sec and a final extension at 70uC for 8 min. Hybridization was performed with the automatic machine (twincubator). After hybridization and washing, strips were removed, fixed on paper and results were interpreted [11]. Each strip of LPA had 27 reaction zones (bands), including six controls (conjugate, amplification, M. tuberculosis complex (TUB), rpoB, katG and inhA controls), eight rpoB wild-type (WT1-WT8) and four mutant probes (rpoB MUT D516V, rpoB MUT H526Y, rpoB MUT H526D, and rpoB MUT S531L), one katG wild-type and two mutant probes (katG MUT S315T1 and katG MUT S315T2), and two inhA wild type and four mutant probes (inhA MUT1 C15T, inhA MUT2 A16G, inhA MUT3A T8C, inhA MUT3B T8A) ( Figure 1). Either missing of wild-type band or the presence of mutant band was taken as an indication of a resistant strain. Incomplete amplification of RIF and/or INH genes was considered as an invalid result. Laboratory staff performing conventional DST was unaware of LPA results, and vice versa. The turnaround time (hrs) was calculated from collection of samples to the availability of results.

Statistical Analysis
Data were analysed by using Stata 11.2. Data were presented as frequency (percentage) and mean (SD). Sensitivity, specificity, negative predictive value, and positive predictive value with 95% confidence intervals were calculated. Categorical variables (like mutation group) were analysed by Chi square/Fisher's exact test. P value less than 0.05 was taken as statistically significant.

Results
A total of 269 previously treated sputum smear-positive patients (172 males and 97 females) were enrolled in the study. Their mean age was 34 (14, SD) yrs. All sputum samples were subjected to AFB microscopy, culture and DST by LJ and LPA methods. The smear examination yielded following bacillary load, 1+ AFB in 62 (23%), 2+ in 109 (41%) and 3+ in 98 (36% Considering the phenotypic conventional Lowenstein-Jensen 1% proportion method as the gold standard, performance of GenoType MTBDRplus LPA in detecting resistance to rifampicin, isoniazid and MDR-TB is detailed in Table 2.
The most frequent mutation found in INH resistant strains was KatG mutation [77/87 (88%)] which occurred more commonly in    (Fig. 1). Isolated inhA gene mutation was found in 4/66 (6%) of MDR-TB strains and 6/21 (28%) of INH monoresistant strains. The turnaround time of line probe assay was 48 hours; whereas, it was 70 days for phenotypic DST (28 days for conventional culture growth and another 42 days for DST).

Discussion
In the recent years, a major emphasis has been given on rapid diagnosis and prompt initiation of accurate treatment of MDR-TB. Accurate and early diagnosis of MDR-TB is highly desirable as it interrupts further transmission of the disease and avoids empirical addition of life-saving drugs and thus amplification of drug resistance and creation of extensively drug resistanttuberculosis (XDR-TB). It also avoids unnecessary cost of administration and occurrence of serious side -effects of second-line anti-tuberculosis drugs in case one is dealing with drug sensitive M. tuberculosis strains.
Present study was conducted in the mycobacteriology laboratory of a tertiary care centre in north India which is accredited for carrying out both DST on solid culture and LPA test. In this study, we evaluated performance of LPA test directly on sputum samples obtained from MDR-TB suspects. Patients were administered MDR-TB treatment based on LPA findings. Laboratory results were compiled later by a physician who was providing care to these patients. Subsequently, genotypic (LPA) and phenotypic (Lowenstein-Jensen proportion method) DST results were compared. We observed that the LPA test results had a good concordance with the conventional DST with a additional advantage of a shorter turnaround time.
In the present study, sensitivity (98%), specificity (99%) for detection of rifampicin resistance and specificity (99%) for detection of isoniazid resistance are in agreement with results of meta-analysis done by Ling et al [12]. However, a slightly higher sensitivity (92%) for detection of isoniazid resistance was observed in our study indicating that most of the mutations conferring INH resistance were found in the gene region that were incorporated in the present LPA strip [13]. Sensitivity (97%) and specificity (100%) for detection of MDR-TB in the present study corroborated with a previously reported study [14] and these findings suggest that performance of LPA is similar to conventional DST in a qualityassured TB laboratory. Sensitivity for detection of RIF resistance    [16]. Additionally, this test can also be useful for systematic surveillance of INH monoresistance in countries with high isoniazid resistance.
In the present study, we encountered a major discrepancy in detecting INH resistance by genotypic method. The LPA test failed to detect INH resistant strains in 7 specimens, suggesting presence of some unidentified mutations in other genomic regions (like ahpC, kasA, furA) were not targeted by the assay used in the present study [17].
Rifampicin resistance is known to be associated with mutations in 81 base pair region (codon 527 to 533) of the rpoB gene [18,19,20]. The finding of dominant mutation for RIF resistance in rpoB S531L [51/71 (72%)], in the present study is similar to a previously published report [21]. One false RIF resistant strain with missing WT8 band was observed with the LPA test.
Distribution of mutations of katG and inhA genes is known to vary in different geographical regions. Frequencies of katG gene, inhA gene and combined katG and inhA gene mutations in the present study [72/87 (83%), 10/87 (11%) and 5/87 (6%) respectively] are within the range of previously reported studies [13,22,23,24]. Finding of frequency of combined mutations of KatG and InhA in the present study is comparable to a recent study from Mumbai [25]. However, as compared to our study authors failed to find an association of katG and inhA genes mutations in INH monoresistant and MDR-TB strains.
Although there is no evidence of any relationship between specific mutation(s) conferring isoniazid resistance and treatment outcome; however, a recent study has reported better treatment outcomes in patients with katG mutations who received high-dose of isoniazid than those who received standard dose of the drug [26], suggesting clinical utility of detection of INH monoresistance pattern with the present LPA test.
The use of recent version of the assay in the present study provided an additional yield of INH resistance due to incorpora-tion of inhA promoter region probes in the test [10/87 (11%); (4/ 66 (6%) in MDR-TB strains and 6/21 (28%) in INH monoresistant strains]. This could not have been detected by the previous version of line probe assay GenoType MTBDR as it lacked the inhA gene probe [27].
The present study highlights limitation of the conventional culture method because of valid LPA results in contamination [4/ 7; (57%)] or negative culture results [3/8; (38%)] and demonstrates superiority of LPA test. Majority of invalid LPA results in the present study were found in sputum specimens with lower bacillary load (1+), or culture negative samples, supporting the recommendation that the direct use of LPA test is not suitable for smear-negative clinical specimens [4,28].
Limitations of Genotype MTBDRplus assay include need for an appropriate infrastructure, adequately trained and skilled laboratory personnel. The test is also not useful in sputum specimens with lower bacillary load and paucibacillary extrapulmonary TB specimens.
It is concluded from the present study that LPA test is highly sensitive and specific for rapid diagnosis of MDR-TB. The test detects resistance patterns with significantly lesser turnaround time as compared to conventional DST method. Additionally, the test also detects monoresistance to isoniazid and rifampicin. More studies with large number of samples are required to validate these preliminary findings, and define the exact place of this test in the diagnostic algorithm for MDR-TB under programmatic settings in high TB burden countries.