TB is increasingly diagnosed using the GeneXpert platform, which can be used for a variety of tests (not just TB). HIV viral load monitoring is required at least annually in patients on ART to detect failure of virologic suppression, however, most HIV VL testing is done centrally. A patient with virologic failure is more likely to get TB.

The investigators wish to see if Xpert done at the clinic results in faster patient TB diagnosis and treatment initiation compared to sending specimens away for central testing. In a different patient group (PLHIV returning for HIV treatment monitoring), the investigators wish to see if POC Xpert HIV-1 viral load (Xpert VL) testing results in faster patient viral load quantification compared to centralised testing. Both POC tests will use the same testing hardware. This polyvalent utility of the GeneXpert system is hitherto uninvestigated in this local setting.

Newly diagnosed pre-ART HIV positive patients will be approached and asked to be a part of this study. Patients will be randomly assigned to Ultra done at the clinic or the normal off-site laboratory TB testing. The time taken for patients to get diagnosed and time-to-treatment will be recorded. We will also do exploratory diagnostic accuracy evaluations including but not limited to, Ultra when done on mouth samples, the new SILVAMP FujiLAM on urine, and host RNA blood signatures for active TB. Additionally, a different group of HIV positive patients (on ART) returning to the clinic for annual follow-up visits will also be asked to join the study. These patients will be randomly selected for either Xpert VL testing done at the clinic or the normal off-site testing. The time taken for patients to receive viral load results will be recorded. Should the patient's viral loads be found to be higher than anticipated and considered by the clinical to indicate a lack of viral suppression, the time taken for patients to have ART regimen adjusted, receive adherence counselling or received HIV drug susceptibility testing will be recorded.

This project will confirm if Ultra TB testing performs well in PLHIV irrespective of symptoms and may produce evidence that supports universal TB testing in this important and vulnerable patient group, including using novel diagnostics on non-traditional specimen types. The investigators will also assess whether POC placement of Ultra and Xpert VL has benefits (e.g., more patients diagnosed for TB or VL monitored during the same day visit).

Sensitive and rapid point-of-care diagnostics should improve TB treatment outcomes, however, tests meeting these criteria have, until recently, been unavailable, especially in PLHIV. PLHIV often have early stage TB disease at the time of ART initiation and paucibacillary sputum. The current frontline test for TB is the Xpert MTB/RIF, which uses the GeneXpert platform, and is deployed primarily at centralised reference laboratories (Clouse et al., 2012; Hanrahan et al., 2015). This approach has two major limitations: 1) the instruments' far-patient placement likely undermines its potential clinical impact; 2) Xpert MTB/RIF has suboptimal sensitivity in PLHIV 3) the GeneXpert platform is primarily used only for TB testing and no other assays such as Xpert HIV-1 Viral Load (VL).

Xpert MTB/RIF has been succeeded by Xpert MTB/RIF Ultra (Xpert Ultra), which promises to increase the speed and sensitivity of TB diagnosis. Although Xpert Ultra will doubtlessly improve the detection of symptomatic patients, its greatest incremental benefit will likely arise in patients with low bacillary load (e.g., unselected HIV-positive patients initiating ART). Thus, Xpert Ultra has the potential to alter how TB is diagnosed in PLHIV by detecting TB before the disease has a chance to progress and before substantial transmission has occurred. The investigators will, in addition to Xpert Ultra on sputum, also perform TB testing using the urinary lateral flow (LF) LAM test, which may detect mycobacteria in asymptomatic PLHIV patients (Lawn et al., 2011). Moreover, investigators will also perform Xpert Ultra and Fujifilm SILVAMP (FujiLAM) on urine in ART initiators. Oral sampling will be done which includes the use of Ultra on tongue swabs and oral washes, inhouse PCR testing MGIT960 liquid culture from PLHIV (ART initiators) as part of preliminary investigation will also be explored, as well the use of blood RNA signatures for TB diagnosis.

These novel test investigations are exploratory, and the results will not yet be used for patient management. This study will address the following research questions: (1) What is the sensitivity and specificity for each approach, overall, and after stratification by viral load/CD4, and (2) What is the proportion of patients that could not expectorate sputum were detected by non-sputum based tests?

The polyvalent utility of the GeneXpert hardware platform is, however, hitherto largely unexplored despite the widespread deployment of machines. HIV VL testing is currently done in South Africa according to the National HIV Treatment guidelines, which advise measuring HIV VL every six months for the first year of treatment and annually thereafter. HIV VL testing involves collection of a blood sample, transporting it to a centralised laboratory for VL quantification, reporting the result back to the clinic, and calling the patient back. If HIV VL is found to be over 1000 genomes per ml, the patient may not be adherent to the prescribed ART or may have drug-resistant HIV and, now that a failure of virologic suppression has been confirmed, the patient may be asked to give a second blood sample for drug susceptibility testing and may have a counselling visit scheduled, in order to improve the patient's adherence. There is, however, a large amount of discretion at local clinics as to what constitutes virologic failure and patients with increased VL may still receive these interventions, even if below the 1000 genomes per ml threshold. The investigators also intend to perform Xpert HIV-1 VL testing and, compared to patients who receive the standard-of-care of centralised VL testing, evaluate the proportion of patients without virologic suppression who are referred to a follow-up intervention (DST and/or adherence counselling).

The investigators propose a study which implements Xpert Ultra (in unselected PLHIV) and Xpert HIV-1 VL (in PLHIV on ART) in Cape Town, South Africa. Furthermore, the GeneXpert platform's polyvalent feasibility, time-to-result and -treatment (Xpert Ultra and LF LAM), and effect on interventions to improve VL (Xpert HIV-1 VL) will be examined.

HIV-positive viral load monitoring patients

Inclusion Criteria:

• Informed consent obtained from patient
• Patient is more than 18 years old
• Patient is HIV positive
• Patient is receiving follow-up ART
• Patient is willing to provide blood specimens for study

Exclusion Criteria:

• Informed consent not obtained from patient
• Patient is less than 18 years old
• Patient is HIV negative or has unknown HIV status
• Patient is coming into clinic for first time ART
• Patient is not willing to provide blood specimens for study

TB patients

Inclusion Criteria:

• Informed consent obtained from patient
• Patient is more than 18 years old
• Patient is HIV positive
• Patient is coming into clinic for first ART visit
• Patient is willing to provide sputum and urine (blood is desirable, but not mandatory) specimens for study

Exclusion Criteria:

• Informed consent not obtained from patient
• Patient is less than 18 years old
• Patient is HIV negative or has unknown HIV status
• Patient is not coming into clinic for first time ART
• Patient has been on TB treatment within the last 60 days
• Patient is not willing to provide sputum and urine specimens for study

Collaborators will have access to de-identified IPD data. Other researchers wishing to access the data will be required to sign a data sharing agreement as per institutional policies.

The data will be available once the findings are published.

Data access will be granted to interested collaborators and parties through sharing of data files or login information for the study's secure RedCap electronic database during the course of patient recruitment.