The optimal diagnosis of active TB involves early and accurate bacteriological confirmation in order to confirm M. tuberculosis infection and to detect potential anti-mycobacterial drug-resistance at baseline. Early diagnosis of HIV-associated TB can be difficult due to the non-specific clinical symptoms, inadequate diagnostic tools available and failure to consider it in the differential diagnosis. A history of past treatment for TB, contact history with a known case of TB, and residence or travel in settings with high endemicity for drug-resistant TB (DR-TB) all suggest an increased risk for resistance and should be ascertained in the clinical history in addition to testing on microbiological specimens.
The diagnosis of M. tuberculosis infection can be established by microscopy, mycobacterial culture or nucleic acid amplification techniques (NAAT), in combination with histological examination of tissue biopsy samples and radiography. Sputum smear microscopy allows diagnosis when there are more than 5000 bacilli/mL of sputum and only detects 40-70% of all TB cases. The sensitivity of sputum microscopy depends on the quality of sputum collection and the preparation and interpretation of microscopy slides. Techniques to improve the yield of sputum samples include sputum induction and broncho-alveolar lavage. HIV-associated TB is more often pauci-bacillary and therefore smear-negative, accounting for 24 to 61% of all pulmonary cases in people with HIV infection. Light microscopy has been more recently replaced in some settings with the more sensitive fluorescence microscopy.
Gold standard bacteriological confirmation requires mycobacterial culture on automated liquid or solid media or both and is highly sensitive, being able to detect 10 bacilli/mL. The limitations of culture are that it requires highly specialised laboratory infrastructure, trained technicians, quality assurance procedures and takes up to 4 to 6 weeks for the result to be reported. Drug-susceptibility testing (DST) should be performed on all culture-positive TB isolates to diagnose DR-TB. This ‘phenotypic test’ evaluates the growth (or metabolic) activity of the M. tuberculosis isolate in the presence of the drug. Drug-resistant isolates of M. tuberculosis are common in many parts of the world and should be considered in people who are likely to have acquired the infection outside of Australia, and in people who have been previously treated for M. tuberculosis infection and have experienced a relapse. Globally, transmission is now considered to be the main route of acquisition, rather than amplification of resistance within an individual.
Molecular or ‘genotypic tests’ undertaken by NAAT can be used to detect drug resistance by identifying mutations in genes of the M. tuberculosis isolate. NAAT has enabled a more rapid and accurate diagnosis of M. tuberculosis infection in biological specimens (increasing the yield in smear-negative pulmonary cases). The Xpert MTB/RIF test is a cartridge-based fully automated platform that detects M. tuberculosis and rifampicin resistance by automated real time polymerase chain reaction (PCR), providing results in less than 2 hours. Rifampicin resistance is a reliable predictor of MDR-TB in high burden settings (MDR-TB prevalence > 10%), however a ‘phenotypic’ DST should be performed if possible. The Xpert MTB/RIF test was endorsed by the WHO in 2010 for use in high-burden countries and is the recommended initial test for TB in people with HIV infection. It increases the detection of smear-negative TB by 67%. The performance is not significantly affected by HIV infection, with a sensitivity of 81% and specificity of 98%. In possible extrapulmonary TB, every effort should be made to obtain samples for microscopy, NAAT, culture and histology. Sensitivity for NAATs is lower in extra-pulmonary samples when compared to sputum. The Xpert MTB/RIF test has good sensitivity (80%) and excellent specificity (> 98%) when performed on cerebrospinal fluid, lymph node aspirates and gastric aspirates, where it can replace smear microscopy.ssay, Xpert MTB/rif Ultra (Xpert Ultra), has been introduced. Xpert Ultra has greater sensitivity for detecting TB than Xpert but at the cost of a slight reduction in specificity. False positive Xpert Ultra results are more common in persons previously treated for TB. Other commercially available molecular techniques are available such as line probe assays (Hain MTBDR, for detection of isoniazid and rifampicin resistance mutations, and Hain Genotype MTBDRsl, for detection of fluoroquinolone and injectable resistance) and many diagnostic laboratories also offer an in-house PCR assay.
Detection of M. tuberculosis liporabinomannan (a component of the cell wall) in urine by lateral flow assay (Alere Determine TB LAM Ag) is a low-cost, point-of-care test that has a high specificity but variable sensitivity depending on the CD4+ T cell count (sensitivity of 66.7% when CD4+ T cell count is <50/μL). It is recommended by WHO for use as a diagnostic test for TB in people with HIV infection who have a CD4+ T cell count <100/μL or those who are seriously ill, as a randomised trial showed a mortality reduction in this group compared with routine care. A newer version (Fujifilm SILVAMP TB LAM) assay has been developed and is under further evaluation, with the initial studies demonstrating a superior sensitivity of 70.4%.
Whole genome sequencing (WGS) and targeted next generation sequencing of M. tuberculosis DNA have great potential for rapid diagnosis of DR-TB through the detection of drug resistance mutations, in addition to providing information for surveillance and public health responses (clustering of strains). DNA sequencing is the most accurate of the molecular tests and is increasingly cheaper and quicker than phenotypic testing. It is possible that DNA sequencing will be able to replace phenotypic testing in the future if automated, affordable platforms and improved understanding of the clinical implication of drug resistance mutations is developed. Some laboratories in Australia provide WGS routinely for culture positive M. tuberculosis isolates. Drug resistance mutations can be classified by how strongly they are associated with phenotypic resistance tests. However, mutations for some drugs are not yet fully known, especially for some of the newer drugs used for DR-TB, and there is still limited understanding of causes of discordance between phenotypic tests and sequencing. Interpretation of results is recommended to be done in consultation with an expert. New diagnostic tests and biomarkers for TB are in development, with several tests in the process of having evidence submitted to the WHO for review prior to a recommendation being made on their use.
The radiological features of TB in people with HIV infection and CD4+ T cell counts >200 /μL are similar to the general population, with a predominance of upper lobe abnormalities, cavitatory disease and the presence of pleural eﬀusions. In immunodeﬁcient persons (CD4+ T cell count <200/μL), mediastinal lymphadenopathy, middle and lower zone infiltrates, non-cavitary disease and extrapulmonary disease are more common. Up to 10% of patients will have a normal chest X-ray. Radiography, including computerised tomography and ultrasound, may be a useful adjunct test in localising and defining the site of extrapulmonary disease.
Screening tests for TB
People with HIV infection in Australia who are from high-burden TB countries, or have travelled to them, should be systematically screened for active TB and LTBI, given their increased risk of having these problems. Screening strategies based on cough only have a low sensitivity for HIV-associated TB. Chest X-ray used in combination with symptom screening has the highest sensitivity for detecting TB and, thus, should be used to exclude active TB before initiating treatment for LTBI. Software for artificial intelligence-based computer aided detection of TB from chest X-rays has been developed and clinical evaluation is ongoing.
Latent tuberculosis infection
The diagnosis of LTBI can be established with the TST or an IGRA, commercially available as QuantiFERON-TB assay and T-SPOT TB assay. Neither assay can differentiate accurately between latent and active TB. The TST is an intradermal injection of tuberculin that stimulates a delayed-type hypersensitivity immune response, mediated by T lymphocytes and macrophages, that results in skin induration, measured by a clinician. False-positive TSTs can be caused by T cell responses to antigens of non-tuberculous mycobacteria (NTM) or Bacillus Calmette-Guerin (BCG), resulting from previous BCG vaccination. The IGRAs measure T-cell responses to M. tuberculosis-specific proteins in whole blood (QuantiFERON-TB Gold In-Tube assay) or by enzyme-linked immunospot assay (T-SPOT TB assay) and do not detect T cells reactive to NTM or BCG. They therefore offer greater specificity than the TST.
As both IGRAs and TSTs measure T-cell responses, they are more likely to be positive in people with HIV infection who have relatively high CD4+ T cell counts than in those with severe immunodeﬁciency. A lower cut-oﬀ for a positive TST is used in the HIV-infected population (5 mm induration rather than 10 mm used in those without HIV infection). As there is no gold standard for mycobacterial detection in LTBI, sensitivity is determined among culture-confirmed active TB cases. The sensitivity of the QFN-TB test is 67% in people with HIV infection and there is no current evidence to suggest superiority of IGRA over TST in the diagnosis of LTBI in people with HIV infection.49
Given the high risk of progression to active TB in individuals with HIV infection, all patients from countries where M. tuberculosis infection is endemic, or who have been exposed to someone with TB, should be tested for LTBI. National guidelines vary in their recommendation of testing algorithm to include one, both or sequential TST and IGRA. The Australian National TB Advisory Committee (NTAC) recommends a TST or IGRA or both, with a diagnosis of LTBI to be made if either test is positive. Patients with a negative TST or IGRA and advanced HIV disease (i.e. CD4+ T cell counts < 200/µL) should have a repeat TST or IGRA after initiation of cART and increase in the CD4+ T cell count to > 200/µL.