GeneXpert MTB/RIF

Xpert MTB/RIF (and the newer Xpert MTB/RIF Ultra) is a cartridge based, automated, real time PCR assay that detects Mycobacterium tuberculosis complex DNA and rifampicin resistance mutations in the rpoB gene (the 81 base pair rifampicin resistance determining region) in about 2 hours directly from sputum and some other specimens.

It requires minimal training and no biosafety cabinet, and is used at the point of care in many countries. Ultra uses two multicopy targets (IS6110 and IS1081) and an improved rpoB melt curve analysis, lowering the limit of detection from about 130 to about 16 colony forming units/mL.

For sputum, sensitivity of Xpert MTB/RIF is about 89% overall (98% in smear positive and 67% in smear negative, culture positive cases), with specificity of 99%.

For Ultra, sensitivity is 88% overall (and 64% to 77% in smear negative, rising by 5 to 17 percentage points compared to the older cartridge, particularly in HIV and children) with specificity of 96% (slightly lower, mostly in people with prior tuberculosis). For rifampicin resistance, sensitivity is about 95% and specificity 98%.

It has become the first line test for presumptive tuberculosis in many programmes and detects about 2 to 3 times more cases than smear in HIV.

Indications. Presumptive pulmonary tuberculosis (as the initial diagnostic test in place of smear). HIV-positive patients with symptoms. Children (sputum, induced sputum, gastric aspirate, nasopharyngeal aspirate, stool). Suspected extrapulmonary tuberculosis (CSF, lymph node aspirate or biopsy, pleural fluid, tissue).

Detection of rifampicin resistance in smear positive or culture positive patients. Contacts of rifampicin resistant tuberculosis. Patients with previous treatment (relapse, failure). Rapid screening in outbreaks. Not for monitoring treatment response.

Specimen. Sputum (minimum 1 mL, preferably 2 to 4 mL), processed within a few days (stable 3 days at ambient temperature up to 35 °C, or 7 to 10 days refrigerated). CSF (centrifuged, pellet at least 1 mL for Ultra and preferably more), tissue and lymph node aspirates, gastric aspirate and stool (processed). The sample is mixed with sample reagent, incubated for 15 minutes, and transferred to the cartridge.

Reference values. Result categories: MTB not detected; MTB detected with rifampicin resistance not detected (semi-quantitative level: very low, low, medium, high); MTB detected with rifampicin resistance detected; MTB detected with rifampicin resistance indeterminate; trace (Ultra only); invalid, error or no result.

Interpretation.

• MTB detected, rifampicin resistance not detected: confirms tuberculosis, and the usual first line regimen starts. Culture and phenotypic DST are still done.
• MTB detected, rifampicin resistance detected: probable rifampicin resistant tuberculosis. Treatment for MDR-TB is started or reviewed pending line probe assay or sequencing for isoniazid and fluoroquinolone resistance. In low prevalence settings, a confirmation test is done as false positive rifampicin resistance can occur (especially with "very low" or "low" bacillary load).
• Ultra "trace" result: indicates a very low number of bacilli. In HIV-positive patients, children and extrapulmonary specimens, trace is treated as positive. In others, clinical judgement is used, with a history of treated tuberculosis in the last 5 years suggesting residual DNA from dead bacilli.
• MTB not detected: reduces but does not exclude tuberculosis (negative predictive value 90% to 99% depending on prevalence). If clinical suspicion is high and the first test negative, further tests (second sample, culture, radiology, histology) are done.
• Indeterminate rifampicin resistance: repeat testing on a new sample.
• Invalid or error: repeat on a new sample (causes include inhibitors, inadequate sample volume, cartridge or power problems).
• Xpert does not differentiate dead from viable bacilli, so it should not be used to monitor treatment or in the first 2 years after treatment completion (false positive from residual DNA).
• Cerebrospinal fluid: sensitivity of Ultra is about 70 to 90% (against culture), higher than the earlier cartridge, but a negative result does not exclude tuberculous meningitis.

Critical values and findings needing immediate action.

• MTB detected in a patient with meningitis, miliary disease or critical illness: start treatment immediately with corticosteroids where indicated, and consider HIV status.
• Rifampicin resistance detected: isolate if infectious, notify public health, start or adjust to an MDR-TB regimen after testing for isoniazid and fluoroquinolone resistance, and screen household contacts.
• MTB detected in a healthcare worker, school child or pregnant woman: urgent treatment and contact tracing.
• Trace result in an HIV-positive patient with advanced disease: manage for tuberculosis, and consider urine lipoarabinomannan and further sampling.
• Repeated indeterminate results with high clinical suspicion: culture and DST in a reference laboratory.

Interferences and pitfalls.

• False positive MTB detection occurs in recently treated patients (DNA from dead bacilli persists for months to years), and false positives of rifampicin resistance occur at low bacillary load because of probe failure.
• Silent mutations in the rpoB gene can be reported as resistance, and some resistance conferring mutations (such as I491F outside the target region) are missed. Phenotypic DST or sequencing resolves discordance.
• Sensitivity is reduced in smear negative samples, children, extrapulmonary specimens (particularly pleural fluid, about 40 to 50%) and HIV. A negative test needs follow up.
• Xpert does not detect isoniazid, fluoroquinolone or other drug resistance, and a separate test (line probe, Xpert MTB/XDR, sequencing) is needed.
• Cartridge reagents must be stored at 2 to 28 °C, and the instrument needs calibration, stable power, temperature control and module checks.
• Sample quality matters: saliva like or insufficient specimens give lower yield, and blood stained sputum inhibits the reaction.
• Xpert detects only the M. tuberculosis complex and not non-tuberculous mycobacteria.
• It should not replace culture where DST is needed, and a positive result does not give the drug susceptibility profile for other drugs.

Special groups.

• Children: stool and nasopharyngeal aspirate improve yield. A negative test does not exclude childhood tuberculosis, and clinical algorithms and radiology are used.
• Pregnancy: no restrictions on testing.
• HIV: tests are especially useful, and urine lipoarabinomannan with Xpert on sputum gives the best yield in advanced disease.
• Previously treated patients: positive results can be residual DNA, so culture and smear are done in parallel.
• Regions: in countries with limited laboratory capacity, Xpert has expanded access, but cost of cartridges, maintenance, power and data systems are constraints.

Follow up tests. Culture and phenotypic DST, line probe assay for isoniazid and fluoroquinolone resistance, Xpert MTB/XDR or targeted sequencing, smear microscopy, chest radiograph or CT, HIV test and CD4 count, urine lipoarabinomannan, CSF analysis, tissue histology, contact screening, liver function tests, and baseline audiometry and ECG for drug resistant regimens.