Assessment of Conventional and Molecular Methods in the Routine Management of Tuberculosis in a High Tuberculosis Burden Setting

.


Introduction
The alarming incidence of pulmonary tuberculosis (PTB) and extrapulmonary TB (EPTB) cases is a global concern.Based on estimation, 1.5 million deaths (1.4-1.6 million) were reported in 2018 due to TB. 1 This problem is further compounded by the emergence of human immunodeficiency virus (HIV)-associated TB infection. 1s per a recent estimate, about 70% of the 0.8 million EPTB cases out of a total of 5.4 million new TB cases were localized in Southeast Asian countries, with India ranking first having a maximum number (about 0.35 million) of EPTB cases. 1 Emerging drug resistance in the form of multidrugresistant TB (MDR-TB) and monoresistant TB is a significant public health problem globally.Among monoresistance in TB, isoniazid (INH) resistance is reported to be the most common.According to the Global TB Report 2018, the prevalence of MDR-TB and INH-monoresistant TB is indicated to be 3.5% and 7.1% in new PTB cases, as well as 18% and 7.9%, respectively in previously treated PTB cases, respectively.1 However, limited information is available on drug resistance in EPTB from a high TB burden country such as India. 2 The reason could be difficulty obtaining specimens and limited laboratories offering culture and drug susceptibility testing (DST) for EPTB samples.
GenoType MTBDR assay, also termed line probe assay (LPA), is a molecular method that detects Mycobacterium tuberculosis complex (MTBC) even in paucibacillary cases but also detects mutations in genes responsible for drug resistance viz.rpoB (for Rifampicin [RIF] resistance), katG (for high-level INH resistance), and inhA (for lowlevels of INH resistance) genes in less than two days.A new molecular method called GeneXpert MTB/RIF Assay detects M. tuberculosis, along with RIF resistance, an indicator of MDR-TB, within three hours.The later assay requires minimal biosafety infrastructure and training, unlike LPA, which is labor-intensive, time-consuming, and technically demanding.The present study sought to compare the performance of molecular methods, namely, GenoType MTBDR plus and GeneXpert with conventional methods, staining and culture in PTB and EPTB samples.

Materials and Methods
A prospective study was performed from July 2017 to June 2018 in the Department of Microbiology at a Tertiary Care Hospital (HAHC) in Delhi, India, after approval by the institutional research and ethics committee with a patient information sheet provided to all patients, and informed consent was obtained from them for this study.

Study Subjects and Sample Size
The sample size was calculated using the standard formula but was later increased to 1 year.A total of 1000 and 412 consecutive new cases with presumptive diagnoses of PTB and EPTB as per revised National TB Control Programme (RNTCP) guidelines constituted the subjects of the present study. 3The sample size for the study was estimated using the following formula: where Z and d represent the standard deviation (1.96) and confidence interval (0.05), respectively.Prevalence was taken at 20% (as per previous studies) But later, it was decided to include all the samples from the study duration (July 2017-June 2018).

Sample Collection and Transportation Pulmonary Tuberculosis Cases
Patients were given a pre-sterilized universal container (100 mL) to collect the deep expectorated sputum sample (3-5 mL) on two occasions (one spot sample under supervision and the other early morning sample).

Extrapulmonary Tuberculosis Cases
These specimens were divided into aseptically collected specimens (sterile) [fluids such as cerebrospinal fluid (CSF), pleural, synovial, ascitic, and tissues (lymph node, tissue biopsies)] and specimens contaminated by normal flora or not collected aseptically [gastric lavage, urine, and pus]. 4All the samples were collected according to RNTCP guidelines 2016. 4ansportation Specimens were transported to the laboratory as soon as possible and, in case of delay, they were refrigerated at 4 °C to inhibit the growth of unwanted micro-organisms. 3,4

Homogenization
It was performed only in aseptically collected EPTB samples using a homogenizer tube. 5

Decontamination
All the clinical samples were decontaminated regardless of the homogenization step.
NALC-NaOH method 6 for pulmonary samples (sputum) and milder decontamination procedure as per RNTCP guidelines were used for EPTB samples. 7All decontaminated PTB and EPTB samples were divided into three aliquots and kept at 2-8 °C.

Smear Microscopy
The first aliquot of the decontaminated PTB or EPTB sample was utilized to prepare two smears (one for Ziehl-Neelsen [ZN] and the other for fluorescent staining).The staining procedure (both ZN and fluorescence staining) and interpretation of results were made according to RNTCP guidelines 2016 7 (Figures 1 and 2).
Processing of Samples for Culture, GeneXpert Assay, and Line Probe Assay All smear-positive and smear-negative PTB and EPTB samples were employed for culture and GeneXpert assay (Figures 1 and 2).

(I) Culture
One loopful of the re-suspended sediment from the first aliquot of the decontaminated sample was used to inoculate the Lowenstein Jensen (LJ) media slant.The solid cultures were read every week (up to 8 weeks) for growth, and any change observed during the incubation period was subjected to ZN staining.Positive ZN staining was interpreted as the specimen positive for acid-fast bacteria (AFB).

(II) GeneXpert
The second aliquot of the decontaminated sample was utilized to perform this assay according to the manufacturer's instructions. 8The results were interpreted as detected/not detected for MTB and detected/not detected/indeterminate for RIF resistance.

(III) Line Probe Assay
The third aliquot of decontaminated and GeneXpert positive PTB and EPTB samples were put up for LPA, which is based on the principle of DNA strip technology involving three steps, 9

Statistics
The data were tabulated in a Microsoft Excel spreadsheet in the form of a master chart and were used for correlation.A one-way analysis of variance (ANOVA) test was utilized to Smear-negative (n=800) (ZN-and FM-) Presumptive PTB cases (n=1000)

p8
LPA - LPA + (n=90) Staining (Fluorescent and ZN) calculate the P value to determine any significant statistical differences between mean ages and was performed in Microsoft Excel.The sensitivity, specificity, and strength of association [kappa (k) value of light-emitting diodefluorescent microscopy (LED-FM), ZN, and GeneXpert] were measured using standard formulae mentioned below using culture positivity as the gold standard. 10here p o and p e denote relative observed agreement among raters and hypothetical probability of chance agreement, respectively.

Results
A total of 1000 consecutive new cases with the presumptive diagnosis of PTB and 412 samples of patients with the presumptive diagnosis of EPTB were enrolled in the study.

Discussion
The lab diagnosis of TB is primarily based on direct microscopy in PTB cases and culture in paucibacillary PTB and EPTB cases.Culture is a more sensitive method than direct microscopy, which detects AFB only when present at least 5000-10 000 bacilli per milliliter of sputum. 11Nevertheless, the latter test is simple and cheap and remains the most commonly used method for diagnosing TB in developing countries such as India for detecting infectious cases and assessing patient response to treatment.In the present study, the smear positivity rates by LED-FM in PTB and EPTB samples were higher (20% and 18.2%) than in ZN microscopy (11% and 9.2%).
A low smear positivity rate of PTB in our study may be due to the earlier presentation of presumptive cases that are more paucibacillary. 12Considering culture as the gold standard for the diagnosis of TB, 10 the sensitivity of LED-FM was higher (82.6%) than the ZN method (47.8%), but specificity was approximately the same (98.7% for LED-FM and 100% for ZN) and in substantial agreement (k = 0.75) with culture in PTB samples in the current study (Table 5).According to a systematic review of 45 relevant studies, FM was found to be, on average, 10% more sensitive with similar specificity compared to conventional microscopy in sputum samples. 13A similar high sensitivity of LED-FM to the ZN method (50.7% vs. 27.5%)with nearly the same specificity (98% vs. 100%) and the kappa value in substantial agreement (k = 0.71) with culture was observed for EPTB samples in our study (Table 5), which is in accordance with the results reported by Adarsh et al 14 (50% vs. 16.6%sensitivity and 100% vs. 100% specificity) and Munshi et al (45.2% vs. 33.9%sensitivity and 88.13% vs. 90.8%specificity). 15The high sensitivity of LED-FM may be attributed to the stronger affinity of auramine to the mycolic acid of tubercle bacilli than carbol fuchsin 16 and a wider field of smear examination (under 40X compared to 100X in conventional light microscopy), and the appearance of AFB as bright rods against a dark background contributing to better detection.These results support its superior diagnostic performance 17 and the target of its implementation by RNTCP in health facilities with higher workloads. 18However, an essential concern regarding LED-FM is false positivity (due to fluorescence by food particles and artifacts in the sample or any impurity in auramine stain).This further suggests ZN staining by LED-FM in all scanty and doubtful cases. 19he World Health Organization (WHO) has recently endorsed the routine use of two molecular methods (i.e., GeneXpert and LPA) for rapid diagnosis and simultaneous screening of MDR-TB in many countries. 20,21MDR-TB is defined as resistance to both RIF and INH.However, RIF resistance is considered by the WHO as an excellent surrogate marker for MDR-TB as mono-resistance to RIF in M. tuberculosis is rare. 22GeneXpert is the critical investigation in RNTCP that detects M. tuberculosis and RIF resistance due to mutation in the 81 base pair core regions of the rpoB gene.
In the present study, the sensitivity of GeneXpert in smearpositive/culture-positive PTB was comparable with smearnegative/culture-positive PTB (95.7% vs. 87.5%)with an overall specificity of 94.15%.In EPTB cases, sensitivity and specificity were 90.5% and 90.1%, respectively (Table 5).][28] The overall performance of GeneXpert was found to be in substantial agreement (k = 0.71) with the gold standard method, namely, solid culture (Table 5).
In the current study, the performance of GeneXpert in detecting EPTB was found to vary concerning different specimens.Maximum sensitivity was observed in CSF, gastric aspirate, urine, and synovial fluid, followed by pus, tissues, and least in the pleural fluid (Table 5).However, the specificity was the same for all types of EPTB samples compared to 81.2% and 66.7% in tissues, as well as 79.5% and 66.7% in CSF in other studies. 29,30Variations in the sensitivity of GeneXpert could be due to differences in patients' population, type, and quality of clinical specimens.The high specificity of GeneXpert could be due to the closed reaction chamber, which reduces crosscontamination and false positivity.Low sensitivity in the pleural fluid and tissues can be attributed to low bacillary load, 31 as GeneXpert requires 131 bacilli/mL compared to culture with a minimum detection limit of 10 bacilli/mL. 32lthough a high specificity gives clinicians confidence to confirm the diagnosis of TB when the GeneXpert test is positive, a negative result cannot rule out the diagnosis of TB because of low sensitivity.Therefore, the clinician should rely on the results of both culture and GeneXpert.
Despite a short turnaround time (less than 3 hours) and minimal training requirement, GeneXpert has a few limitations such as the limited shelf-life of the cartridges, restrictions in operating temperature and humidity, continuous electricity supply, unknown long-term robustness, and periodic calibration of the machine. 33ur result indicates 37% of EPTB cases and 63% of PTB cases among all TB cases.This is in line with a study reported from Delhi in which the incidence of EPTB and PTB was 30.8% and 69.2%, respectively. 34The most caseload was found in the 20-40 year-age group (Table 4) in both PB and EPTB.Similar results were reported by Bagchi et al and Sharma et al. 34,35 A higher proportion of females were affected by EPTB, while males were more involved with PTB in the current study.However, as per reports from Delhi, males predominate in both PTB and EPTB cases. 34][38] According to the WHO global TB report, the prevalence of MDR-TB in new and previously treated cases was 2.5-2.8% and 14-17%, respectively. 39As a high TB burden country, India is also a hotspot region for MDR-TB infection. 40A 12.9% prevalence of MDR-TB was reported from North India, with Delhi at 9.55%. 34In the present study, a 6.7% prevalence of MDR-TB due to rpoB mutation probe 1 (mutation in D516V) and KatG mutation probe 1 (transformation in S315T1) was observed only in PTB cases.To date, the primary focus of drug-resistant TB in India has been mainly on RIF resistance.According to the first national TB drug resistance survey in India, INH monoresistance was the highest (4% in new cases and 8% in previously treated cases) 36 and the single most common anti-TB drug resistance due to two genes (i.e., KatG and inhA) 37 unlike RIF resistance (rpoB gene), but it has not received much attention yet.This could be due to complex diagnostic procedures using molecular techniques and their uncertain clinical implications.A 3.3% prevalence of INH monoresistant TB was found in PTB cases responsible for KatG mutation probe 1 (mutation in S315T1).More than 90% of INH monoresistance in India is due to mutation in the KatG gene associated with high-level resistance and poor treatment outcomes, 38 and mutations in both KatG and inhA genes responsible for INH monoresistance lead to higher chances for the development of MDR-TB. 35mong DR-TB cases in PTB samples, 75% were MDR-TB, and 25% were INH-monoresistant TB.This percentage of INH-monoresistance should be noticed and should be seriously taken immediately.INH-monoresistant cases will be missed if not precisely looked for.This will result in misdiagnosis, leading to an increased risk of treatment failure or relapse and a greater propensity of progression toward MDR-TB.
A significant hurdle in controlling the INH monoresistance problem in India is the use of cartridgebased nucleic acid amplification test (CB-NAAT) in national programs, which does not identify it.As per RNTCP, if CB-NAAT does not detect RIF resistance, the patient should be looked at explicitly for INH monoresistance by LPA and treated accordingly.However, INH testing is still a severe challenge to India.The reason is a limitation of tools such as LPA and liquid DST to reference or centralized laboratories.There are still large numbers of INH-monoresistant patients who are misdiagnosed currently and mismanaged subsequently.Therefore, the promotion of INH testing in all RIF-sensitive cases or, in general, universal DST on all MTB isolates, followed by individualized therapy, should be strictly performed to control the overall drug resistance to fulfill the ambitious goal of eliminating TB by 2025.

Conclusion
With the increasing prevalence of TB in India, the mainstay of its diagnosis, which is smear microscopy, does not detect drug resistance.MDR-TB is resistant to both RIF and INH.Novel molecular assays such as GeneXpert detect only RIF resistance, while only a few of them, including LPA, detect INH resistance, which is quite common and can lead to therapy failure or rapid progress to MDR-TB.India has set an ambitious goal of eliminating TB by 2025.Limiting the spread of TB, especially DR-TB, by fast diagnostic methods can be a powerful tool to achieve this goal and should not be considered an additional cost.This may be achieved by promoting universal DST for all positive cases.Finally, the physicians should be wide-awake in promptly identifying and treating INH-monoresistant TB accordingly.

Table 1 .
Staining Results by ZN and LED-FM

Table 5 .
Sensitivity, Specificity, and Kappa Value of FM and GeneXpert With Culture as the Gold Standard in PTB and EPTB Cases