A clinical trial conducted in Tanzania and Mozambique and published in The Lancet Infectious Diseases concludes that expanding the use of molecular diagnostic tests on urine and stool samples, in addition to sputum, to detect tuberculosis in hospitalised people living with HIV does not improve early treatment initiation nor reduce short-term mortality.
The study, led by the Barcelona Institute for Global Health (ISGlobal), a centre supported by ”la Caixa” Foundation, together with FIND (Foundation of Innovative Diagnostics, Switzerland), aimed to assess the integration of new molecular tests into current WHO-recommended diagnostic strategies.
Tuberculosis: a complex diagnosis with high mortality
Tuberculosis remains the leading cause of death among people living with HIV. Diagnosing the disease in this population is particularly challenging: symptoms are often non-specific, the disease may present outside the lungs, and many patients are unable to produce sputum, a key sample for diagnosis. In addition, the bacterial load in respiratory samples from patients with HIV is often low, reducing the likelihood of microbiological confirmation.
In response to these limitations, recent years have seen the development of alternative diagnostic approaches based on non-invasive samples such as urine and stool, with the aim of improving case detection and accelerating treatment initiation.
Intensified TB screening in hospitalised patients with HIV
The EXULTANT trial, funded by the EDCTP, is a multicentre randomised study conducted in 11 hospitals across Tanzania and Mozambique. It included 1,172 adults living with HIV admitted to hospital, regardless of whether they showed symptoms of tuberculosis. The primary objective was to assess the proportion of patients with microbiologically confirmed tuberculosis who were able to start treatment within the first 72 hours after admission. Time to diagnosis and treatment initiation, as well as mortality at eight weeks, were also evaluated.
“Participants were assigned to two groups: the intervention group received expanded screening with Xpert MTB/RIF Ultra molecular tests on sputum, urine and stool samples, along with a urine LF-LAM test,” explains Marta Cossa, researcher at CISM and first author of the study. “The control group followed the standard WHO-recommended protocol, based on symptom assessment and selective use of diagnostic tests.”
More diagnostic testing does not speed up treatment or reduce mortality
The results showed that 16.0% of patients in the expanded screening group started treatment within the first 72 hours, compared with 15.3% in the standard group, with no statistically significant difference. Likewise, no significant differences were observed in eight-week mortality between the intervention group (25.8%) and the control group (28.8%).
Overall, the expanded strategy did not increase the number of confirmed diagnoses nor shorten the time to treatment initiation. Although molecular testing of urine and stool samples identified some additional TB cases, the urine LF-LAM test detected the highest number. However, this improved diagnostic yield did not translate into earlier treatment or reduced eight-week mortality, raising questions about the test’s clinical impact and highlighting the need for further evaluation of its accuracy.
“Approximately half of the patients were able to provide sputum samples, which reinforces the role of urine and stool as alternative diagnostic samples,” explains Alberto García-Basteiro, ISGlobal researcher and principal investigator of the EXULTANT trial. “Future studies should focus on optimising diagnostic algorithms and improving the implementation of available tests in hospitalised people living with HIV,” he concludes.
Reference
Cossa, M., Ndege, R., Meggi, B., Mangu, C., Leukes, V., Sabi, I., Khosa, C., Kivuma, B., Nguenha, D., Riess, F., Fernández-Escobar, C., de la Torre-Perez, L., Manjate, P., Weisser, M., Machiana, A., Zachariah, C., Tagliani, E., Ehrlich, J., Buech, J., … Garcia-Basteiro, A. L. (2026). Expanding Xpert MTB/RIF Ultra and lateral flow urine lipoarabinomannan testing for diagnosis of tuberculosis among adults living with HIV admitted to hospitals in Tanzania and Mozambique (EXULTANT): A randomised controlled trial. The Lancet Infectious Diseases, 26
