Evaluation of molecular diagnostic test for detection of adult pulmonary tuberculosis: A generic protocol

Diagnostic tests:

• (i) Reference test: MGIT culture (gold standard)/reference standard

• (ii) Index test: new NAAT test under evaluation

• (iii) Comparator: standard NTEP-approved NAAT test (Xpert MTB/RIF used currently in the TB control programme in India)

As the index test would be a NAAT-based molecular diagnostic test, it would be necessary to compare the performance of a NAAT test with another standard NAAT-based test. Since Xpert MTB/RIF is the test in use under ‘National TB Elimination Programme’ (NTEP) and WHO-approved, Xpert MTB/RIF would be used as a comparator in such a study.

Index tests: The new molecular diagnostic tests using open RT-PCR-based kits needs to be compatible with various commercially available RT-PCR systems (e.g. BIORAD CFX96, ABI7600, Roche LightCycler® 480, etc). The index-test kit needs to include gene-specific probes for detection of amplified target DNA sequences of MTB complex. The target genes can be identified by a signal emitted by the reporter dyes or fluorophores [e.g. Texas Red/Cy5, FAM (5-carboxyfluorescein)/HEX (phosphoramidite)] attached to the probes. After extraction of DNA from the sputum samples, 5 to 10 μl of the extracted DNA template needs to be added to the reaction mix and run on an RT-PCR machine for up to 40 cycles or as specified by the manufacturer.

Enrolment: Spectrum bias can be avoided by enrolling a consecutive series of individuals adequately representing the population in which the test will be used. All eligible individuals presenting to the hospital OPD or DMC or clinic of the selected site are to be screened for TB to identify presumptive TB cases. After obtaining written informed consent from such individuals, demographic details of the individuals including their name, age, gender, address, contact details, prior history of TB treatment, clinical history, diagnosis of HIV and radiological findings should be documented in a case record form as per the NTEP norms. All screening evaluations to be completed and reviewed to confirm that potential participants met all the eligibility criteria.

Sample size: The anticipated sensitivity of an index test needs to be kept at 90 per cent and specificity 99 per cent. The aim of the study would be to estimate the sensitivity with an absolute precision of 5 per cent (L=0.05) on either side with a confidence level of 95 per cent and the specificity with an absolute precision of one per cent (L=0.01). A higher precision for specificity would be required to minimize false positivity. The formulas used for calculating sample size for estimating sensitivity and specificity are as follows:

where,

z_1−α/2 = 1.96 for 95 per cent level of confidence

S_N = anticipated sensitivity

L = Absolute precision as desired on either side (half-width of CI) of sensitivity (margin of error)

Prevalence = Anticipated prevalence of the disease

where,

z_1-α/2 = 1.96 for 95 per cent level of confidence

S_P = Anticipated specificity,

L = Absolute precision as desired on either side (half-width of CI) of specificity (margin of error)

Prevalence = Anticipated prevalence of the disease

The prevalence of culture positives among presumptive cases in hospital setting has been reported to be 24 per cent11 (Prevalence = 0.24). With this assumption, the minimum sample size requirement has been calculated as 577 for estimating sensitivity and 501 for estimating specificity with the required precision. The loss due to indeterminate samples is expected to be five per cent. Thus, the sample size required is as follows:

• (i) For sensitivity: 608

• (ii) For specificity: 528

• (iii) The higher of these two is 608.

Thus, a total of 610 consecutive cases meeting the inclusion and exclusion criteria would be enrolled and their sputum samples would be tested as per the algorithm as shown in the Figure. At the expected rate of 24 per cent culture positivity among presumptive TB cases in hospital OPD-setting, this number of enrolled cases would be expected to provide nearly 146 positives and 464 negatives for MTB by the gold standard culture. These can be approximated to 150 and 470, respectively. Enrolment would be continued till the required number of participants are covered.

Close

Figure

Schematic flow through of the proposed protocol. *Screening: Medical history & clinical examination as per NTEP guidelines. #Comparator: GeneXpert as recommended by NTEP. $Index Test: New NAAT Test under evaluation. Storage: Sites to process samples for additional testing and storage as per their routine practice (1-MGIT/1-LJ as per practice). One positive culture and 2 decontaminated samples per patient will be stored at -80°C for later use, if necessary 2 DNA samples will be stored at -20ºC for resolution of discrepant results. PTB, pulmonary tuberculosis; ATT, anti-tuberculosis treatment; MGIT, mycobacteria growth indicator tube; MTB, Mycobacterium tuberculosis; RIF, rifampicin; NAAT, nucleic acid amplification test.

Export to PPT

Sample collection, processing and storage: Two sputum samples, each of minimum 3 ml, to be collected from each presumptive TB case: one spot and one morning sample. Chances to have lesser bacterial load in spot sample would prevail. Since morning sputum sample carries higher bacterial load, two samples are usually collected to have sufficient sputum volume and bacterial load. Especially in a validation study involving multiple tests, it would be preferable to collect two samples for processing culture and molecular tests.

Sputum samples collected from each individual with presumptive TB need to be processed as per the algorithm depicted (Figure) under standard biosafety conditions. Briefly, the sputum samples need to be partially homogenized using sterile beads. Approximately, 1 ml from each spot and morning sample to be pooled under sterile conditions. From the pooled sample, one ml may be used for molecular testing following the manufacturer’s instructions for both standard NAAT (Xpert MTB/RIF) and index test(s).

The remaining portion of each sputum specimen should be used for direct smears and decontamination by N-acetyl-l-cysteine–sodium hydroxide (NaLC-NaOH) method12. The resulting deposit is to be used for concentrated smear preparation and inoculation into two MGIT960 tubes. Lowenstein–Jensen (LJ) slopes may also be used for culture, only for storage or retesting purpose (in case of contamination of all MGIT tubes). All positive cultures can be identified using the rapid immunochromatographic test (ICT) within five days of instrument positivity; interpretation of the results to be done as per the manufacturer’s instructions. For each paticipant, one positive culture and two decontaminated samples will be stored at −80°C for potential future use if necessary. The extracted DNA may also be stored at −20°C until the end of the study to resolve any discrepant results.

Laboratory procedures: All conventional test procedures for smear, culture (solid and liquid) and Xpert MTB/RIF will be performed as per the NTEP national laboratory guidelines12 and laboratory manual of the Indian Council of Medical Research (ICMR)-NIRT13. The standard operating procedure (SOP) for index test(s) provided by the manufacturer(s) including the use of positive and negative controls need to be followed. All procedures for preparation of media, reagents, washing, decontamination, disposal and storage to be performed according to the SOP of ICMR-NIRT.

Tests: (i) Smear microscopy: one direct smear and one concentrated sputum smear

• (ii) MGIT culture: two tubes per specimen – a total of four MGIT tubes for two specimens per participant

• (iii) Identification of culture: rapid immunochromatographic test (ICT) of MGIT culture

• (iv) Xpert MTB/RIF: one test per participant.

Quality control (QC) measures: All sites should ensure high quality of laboratory procedures, data recording and documentation without any deviation from the protocol. All the study sites should participate in internal quality control (IQC) as well as external quality assurance (EQA) for all methods. The standard manuals of the global laboratory initiative14 and the WHO practical manual15 to be adapted to the local laboratory requirements.

Culture: Positive (reference strain H37Rv or H37Ra) and negative controls for MGIT and LJ cultures need to be tested as per the NTEP guidelines. It is required that MGIT time to the detection QC for MTB reference strain is performed every month/new lot of reagents/machine service. Sterility and performance testing of culture media should be performed with every new batch or lot.

Drug sensitivity testing (DST): The standard American Type Culture Collection strains will be used for each drug as reference control. QC will be performed everytime with a new batch of drugs prepared, after servicing of the instrument and after long gap of setting up DST.

Smear: A positive (MTB H37Rv) control smear and a negative control (e.g. Escherichia coli) smear should be included in each batch of tests daily and with new lots of reagents.

ICT Identification of Mycobacterium tuberculosis (MTB) complex: Culture of MTB reference strain in MGIT broth could be used as a positive control. Culture of Mycobacteria other than TB (e.g. a well-characterized strain of Mycobacterium avium complex/Mycobacterium kansasii) in MGIT broth could serve as a negative control.

Molecular diagnostics: For molecular diagnostics, IQC would include a control supplied by the manufacturer. EQA for molecular diagnostics is available in India and offered by the Central TB Division through its national reference laboratories (https://www.naateqa.in/Presentation). EQA programme for Xpert MTB/RIF would contain at least five samples to test the reproducibility, repeatability and accuracy of the assay and need to be carried out at least once a year by the participating laboratories. Each panel for Xpert MTB/RIF consists of (i) MTB not detected, (ii) MTB detected – low, (iii) MTB detected – medium and (iv) MTB detected – high. All the laboratories participating in the study would be part of the EQA programme.

Data analysis plan: If the index test produces an error or indeterminate results, one repeat may be allowed. Samples yielding both contaminated and NTM cultures would be excluded from analysis. The diagnostic accuracy of the index test with reference to culture will be estimated based on the primary analysis (Table I). Interpretation of smear and culture results will be based on the definitions given in Supplementary Tables I and II.

Only absolute positives and negatives by culture are to be considered for primary analysis. Ambiguous culture results (only one culture positive and/or late growth without sign of contamination) should not be considered. In order to resolve discrepancies due to experimental errors, secondary analysis may be performed as per Table II using a comprehensive reference standard716 (demonstration of MTB by standard NAAT or smear).

Listing of constraints such as number of errors or indeterminates by the index test(s), procedural/lab specific/instrument specific/site specific limitations of the index test(s) need to be documented (e.g. requirement of excessive consumables or exclusive instruments /dedicated man power etc). These constraints would be given due consideration while making policy recommendations. Level at which the index test may be employed, e.g. DMC, TB unit and district TB centre or reference lab (IRL/NRL) would also be evaluated based on the constraints such as steps required for extraction and amplification, requirements of additional consumables, biosafety level, skilled staff and time taken.

Statistical analysis: The index molecular test needs to be evaluated for its specificity, sensitivity, accuracy positive as well as negative predictive value. A 95 per cent confidence interval need to be calculated for each of the above indicators as given below:

95 % CI for sensitivity = sensitivity ± 1.96 (SE_sensitivity)

95 % CI for specificity = specificity ± 1.96 (SE_specificity)

The sensitivity and specificity calculation framework to be used with reference to MGIT culture is given below as a Table III.

The positive and negative predictive values will be calculated with reference to MGIT culture as given below:

% Positive predictive value (PPV) of index NAAT test in presumptive TB cases:

% Negative predicted value (NPV) of index NAAT test in presumptive TB cases:

Descriptive statistics on participant characteristics and estimates of diagnostic accuracy should be reported separately for relevant subgroups serving as a proxy for disease spectrum or severity. The acceptable kappa coefficient range would be 0.81-1, indicating almost perfect agreement between the index test and the standard test. Standards for Reporting of Diagnostic Accuracy Studies guidelines17 would be followed for reporting the findings.

Ethics and dissemination: The study should be compliant with the ICMR-National Ethical Guidelines for Biomedical and Health Research involving Human Participants. The institutional ethics committee of each participating site should be intimated about the study for necessary approval prior to initiating the study.

• (i) Written informed consent should be obtained from the participants prior to the study.

• (ii) Two sputum specimens need to be collected from presumptive TB cases as per the algorithm (Figure). No additional sputum or any other specimens will be collected. Probability of harm or discomfort anticipated in the research is expected to be nil or very unlikely.

• (iii) The confidentiality of research participants should be ensured by encrypting the personal identifiers of the participants.

• (iv) Dignity of research participants will be prioritized.

• (v) No compensation to be provided to the participants since there would be no additional cost or travel involved in sample collection for the study. Participants to be compensated for travel and time only if they are asked to pay additional visits exclusively for the sake of the study and not during regular treatment visits.

• (vi) The investigators should not have any conflicts of interest (financial or non-financial). Affiliations in any forms and sources of funding must be declared in advance.

• (vii) The findings of the study should be made accessible through reports.

• (viii) The anti-tuberculosis treatment (ATT) decisions will not be made based on the result of the index test under evaluation, rather on the basis of routine clinical and laboratory tests only (smear, solid/liquid culture, standard NAAT results and clinical workup).

• (ix) Follow-up visits will be required for a few discrepant cases to exclude TB as per NTEP-guideline.

• (x) There should be no delay in reporting and participants should be given the best possible diagnostic investigations.

• (xi) Sputum samples collected should not be stored for any future research. The stored pellets will only be needed to resolve discrepant cases. One positive culture and two DNA samples per participants may be stored at −80°C for later usage.

• (xii) At all the sites all study participants should be followed up till the final diagnosis is made and an individual is initiated on appropriate treatment as per the NTEP norms. Those found to be MTB positive by standard NAAT test may be put on ATT by a medical officer of the study site as per the NTEP guidelines.