Newer and novel sputum versus non-sputum-based tools for the diagnosis of active tuberculosis in different patient sub-populations

Esmail, Aliasgar

Newer and novel sputum versus non-sputum-based tools for the diagnosis of active tuberculosis in different patient sub-populations

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Medicine

Faculty

Faculty of Health Sciences

Abstract

Background Pulmonary tuberculosis (PTB) has a spectrum of presentation ranging from sub clinical/minimally symptomatic disease on one end, usually in community-based patients, to overt symptomatic TB with hospitalisation on the severe end of the spectrum. This spectrum of sub-clinical and clinical presentation of TB along with the site of TB (pulmonary vs. Extra-pulmonary TB) generate a number of sub-populations. Current diagnostic tools are not a ‘one size fits all' and have important differential limitations in these patient sub-populations including those with pauci-bacillary disease (e.g., in HIV-infected patients, minimally symptomatic patients with low burden disease in the community, and in patients with extra-pulmonary TB). Thus, the overarching objective of this PhD was to evaluate the diagnostic performance of frontline tests, including the newer and more sensitive Xpert Ultra and urine LAM in key active TB patient sub-populations (wherein data are limited). The main themes and sub-populations included in my thesis are as follows: 1) Hospitalised HIV-Infected patients (patients in the severe end of PTB spectrum): To determine the optimal and most cost-efficient diagnostic strategy incorporating sputum and LAM (lipoarabinomannan) for the detection of TB (chapter 1). 2) Smear-negative TB: Independent confirmation on the performance of the more sensitive Xpert Ultra (Cepheid MTB/RIF Ultra) in sputum archived samples (chapter 2). 3) Community-based minimally symptomatic patients (patient in the non-severe end of the PTB spectrum): Evaluation of point-of-care (POC) sputum Xpert Ultra for detection of TB in minimally symptomatic/at-risk population for TB and its impact in detecting potentially infectious patients (chapter 3). 4) Extra-pulmonary TB (TB serositis): Evaluation on the performance of conventional diagnostic tools, including Xpert Ultra in tuberculous pericarditis (chapter 4) and pleuritis (chapter 5) compared to a novel immunodiagnostic tool (IRISA-TB). Methods (sub-populations are underlined) In Chapter 1, I describe an evaluation of an algorithm that describes the optimal & cost efficient strategy to combine sputum GeneXpert and urine LAM (tests that are recommended by the WHO) in a hospitalized HIV-infected sub-population: This comprised a post-hoc analysis of 561 HIV-infected sputum-expectorating patients that was part of a larger parent trial. 5 different diagnostic strategies using sputum culture as a reference standard were explored (Xpert alone, LAM alone, sequential Xpert followed by LAM and vice versa [LAM in Xpert-negative patients and Xpert in LAM-negative patients], and both tests concurrently [LAM + Xpert]). A cost-consequence analysis was also performed. In Chapter 2, I evaluated the performance of GeneXpert Ultra in 272 selected and well characterized archived sputum samples including 104 patients with smear negative TB and 102 non-TB with a history of previous TB to accentuate and evaluate the clinical significance of trace readouts (i.e., a readout that usually corresponds to the detection of a very small amount of TB DNA). Assay-specific limit-of-detection (LOD) experiments were conducted using serial dilutions of Mycobacterium tuberculosis H37Rv to confirm the assay's detection threshold. In In Chapter 3, I evaluated the diagnostic performance of point-of-care (POC) Xpert Ultra for the detection of TB in minimally symptomatic community-based sub-population: 5,274 participants were rapidly screened to enrol 584 patients with suspected pulmonary TB from peri-urban high burden communities of Cape Town, South Africa. The utility of POC-Xpert Ultra in detecting likely infectious patients was also specifically evaluated. In Chapter 4, I described the diagnostic performance of pericardial fluid unstimulated interferon gamma (measured using the novel IRISA-TBTM test) and compared it to other same-day tests including Xpert Ultra and adenosine deaminase (ADA) in 99 South African patients with suspected pericardial TB using a composite reference standard including pericardial fluid, pericardial tissue TB culture, pericardial tissue histopathology and response to TB treatment. Similarly, In Chapter 5, I described the diagnostic performance of pleural fluid unstimulated interferon gamma (measured using the novel IRISA-TBTM test) and compared to other same-day tests including Xpert Ultra and adenosine deaminase (ADA) in 207 individuals from Cape Town, South Africa and Vellore, India. A composite reference standard including pleural fluid, pleural tissue TB culture, pericardial tissue histopathology and response to TB treatment was used to define the composite reference standard for TB. Results: Chapter 1: In the HIV-infected hospitalised patient sub-population, the incremental yield of LAM over Xpert was 29.6% (45/152) and that of Xpert over LAM was 75% (84/11). The incremental yield of LAM increased with decreasing CD4 count. The costs per TB case diagnosed were similar for the sequential and concurrent strategies ($1,617 to $1,626). Chapter 2: Xpert Ultra had a lower sputum-specific LOD compared to that of the G4 version of Xpert MTB/RIF (9 vs. 184 cfu/mL). 94% in both forms of EPTB in chapters 4 & 5. Conclusions: Overall, the performance of currently available diagnostic tools varied in different subpopulations due to differential mycobacterial load, the compartment interrogated, patient phenotype (e.g. HIV-infected), testing strategy, and clinical context (e.g. hospitalized versus community-based). Main findings from each chapter as summarized as follows: Main conclusion in chapter 1: Xpert-Ultra had a considerably lower limit-of-detection compared to older Xpert-MTB/RIF assay, however, trace-readouts in the context of previous TB, reduced the specificity of the assay slight (<5%). A third of the trace results (3/9) were likely false positive. Main conclusion in chapter 2: In sputum-expectorating hospitalized patients with advanced HIV and access to both Xpert and LAM, concurrent testing with sputum Xpert and urine LAM may be the best and most cost-effective strategy for diagnosing TB in this sub-population. Main conclusion in chapter 3: POC-Xpert Ultra, when used as part of a community-based ACF, missed ~50% of patients with culture-positive TB. However, Xpert detected almost all likely infectious cases. Main conclusions in chapters 4 & 5: Conventional TB diagnostic tests performed poorly in pericardial and pleural TB. IRISA-TBTM, a novel same-day immunodiagnostic test, outperformed Xpert Ultra for the diagnosis of pericardial and pleural TB. These data have implications for the clinical utility of newer diagnostic tools in patient sub-populations from TB and HIV endemic settings.