Molecular (or genotypic) tests can be used to diagnose TB through the amplification of nucleic acids (DNA or RNA). They are also used to detect drug resistance through identifying genetic mutations (drug-resistant alleles) in the bacterium responsible (genotypic DST). Different assays and platforms have been developed.
3.4.1 Automated real time PCR (Xpert MTB/RIF)
This test can diagnose TB and resistance to rifampicin. In contrast to other techniques (in vitro culture, DST and conventional molecular techniques) the Xpert MTB/RIF can be used in peripheral laboratories and does not require sophisticated equipment or highly-skilled personnel10,11,12,13.
The test is based on real-time PCR, targeting specific nucleic acid sequences in the M. tuberculosis complex genome, while also simultaneously providing information about the most common mutations related to rifampicin resistance.
It is a highly automated test (only 3 manual steps required), which is run in a closed system with one cartridge per sample. Thus, it is less prone to contamination than other PCR-based tests. Each instrument can process 4 samples at one time, with a processing time of just under 2 hours. Higher capacity machines are available. See Appendix 3 for more information on Xpert MTB/RIF instruments.
The performances of this test are almost similar to that of the culture. Published results have shown that for PTB detection, the assay has sensitivities of 98% for smear-positive, culture-positive samples, and 72% for smear-negative, culture-positive samples (sensitivity can reach 90% if the test is repeated 3 times)14,15.
The test Xpert MTB/RIF also has good sensitivity (80%) and excellent specificity (> 98%) when performed on cerebrospinal fluid, lymph node material and gastric fluid16.
Because of its excellent performance, its quick turn around time and its ease of use, this test should be used as an initial diagnostic test in HIV-infected patients and when multidrug-resistant TB (MDR-TB) or TB meningitis are suspected, in both adults and children.
It can also be used for diagnosis of lymph node TB. As the sensitivity of the Xpert test in pleural fluid is low, its use is not recommended.
The sensitivity for the detection of rifampicin resistance compared with conventional DST on culture is 97.6%. The test has a high negative predictive value, therefore, non rifampicin resistant results can be considered to be true susceptible.
In populations where the prevalence of MDR-TB is below 10%, the positive predictive value is below 85% (Appendix 3). Therefore when a Xpert with RIF positive results is found, the test should be immediately repeated in order to rule out possible labelling or clerical errors. If the second Xpert MTB/RIF test does not show rifampicin resistance, the patient can be considered has having a susceptible TB. If the result of the second Xpert MTB/RIF test also shows rifampicin resistance, it should be confirmed by a phenotypic DST or a different genotypic DST method.
Xpert MTB/RIF does not eliminate the need for conventional microscopy, culture and DST, which are required to monitor treatment progress and to detect resistance to drugs other than rifampicin.
3.4.2 Line probe assays (LPA)
To date no fully automated LPA exist. These molecular tests can only be performed by specialized laboratories with strict quality assurance procedures in place.
There are a number of different molecular assays available:
– Conventional Nucleic Acid Amplification (NAA) amplifies M. tuberculosis-specific nucleic acid sequences with a nucleic acid probe, enabling direct detection of the bacillus. The current NAA tests available show a lower sensitivity than culture and therefore, are not recommended for the diagnosis of TB. They are also too labour-intensive to be implemented for routine diagnosis in most laboratories.
– Two molecular techniques are commercially available:
• Hain assays: GenoType® MTBDRplus assay and GenoType®MTBDRsl (Hain Lifescience GmbH, Nehren, Germany). The GenoType® MTBDRplus assay has been shown to be good at detecting rifampicin resistance but less so for isoniazid resistance among smearpositive patients (sensitivity and specificity values for rifampicin and isoniazid were 95.3% and 95.5% and, 89.9 and 87.1%, respectively)17,18. The GenoType®MTBDRsl assay can detect resistance to fluoroquinolones and injectables drugs with a good specificity but a lower specificity (85% for fluoroquinolones and 43 to 84% for injectables)19.
• The INNO-LiPA Rif. TB® line probe assay (Innogenetics, Belgium)20.
The GenoType® MTBDRplus assay can identify mutations on the KatG or on the InhA genes:
– Mutation on KatG gene corresponds to resistance to high-dose isoniazid;
– Mutation on InhA gene corresponds to resistance to both isoniazid and ethionamide, but not necessarily to high-dose isoniazid.
The GenoType®MTBDRsl assay can be used as a triage test on smear-positive patients to guide the initial treatment in extensively drug-resistant TB (XDR-TB) suspects while awaiting confirmatory results from conventional phenotypic testing. However, LPA assays cannot be used as replacement tests for conventional phenotypic second-line anti-TB DST.
These molecular methods have the advantage of giving fast results, within a few hours, for smear-positive patients (referred to as direct testing, because the sputum can be directly tested). For smear negative patients, a primary culture is needed prior to testing (referred to as indirect testing because a culture first has to be grown from the patient’s sputum).
In order to benefit from the short turn around time of these tests, good logistical support is required for sample transportation to the reference laboratory with timely return of results. The main constraints remain the high cost, high infrastructure requirements, high level of technical training and the risk of cross-contamination.