Novel FujiLAM assay to detect tuberculosis in HIV-positive ambulatory patients in four African countries: a diagnostic accuracy study

BACKGROUND: Development of rapid biomarker-based tests that can diagnose tuberculosis using non-sputum samples is a priority for tuberculosis control. We aimed to compare the diagnostic accuracy of the novel Fujifilm SILVAMP TB LAM (FujiLAM) assay with the WHO-recommended Alere Determine TB-LAM Ag test (AlereLAM) using urine samples from HIV-positive patients. METHODS: We did a diagnostic accuracy study at five outpatient public health facilities in Uganda, Kenya, Mozambique, and South Africa. Eligible patients were ambulatory HIV-positive individuals (aged ≥15 years) with symptoms of tuberculosis irrespective of their CD4 T-cell count (group 1), and asymptomatic patients with advanced HIV disease (CD4 count <200 cells per μL, or HIV clinical stage 3 or 4; group 2). All participants underwent clinical examination, chest x-ray, and blood sampling, and were requested to provide a fresh urine sample, and two sputum samples. FujiLAM and AlereLAM urine assays, Xpert MTB/RIF Ultra assay on sputum or urine, sputum culture for Mycobacterium tuberculosis, and CD4 count were systematically carried out for all patients. Sensitivity and specificity of FujiLAM and AlereLAM were evaluated against microbiological and composite reference standards. FINDINGS: Between Aug 24, 2020 and Sept 21, 2021, 1575 patients (823 [52·3%] women) were included in the study: 1031 patients in group 1 and 544 patients in group 2. Tuberculosis was microbiologically confirmed in 96 (9·4%) of 1022 patients in group 1 and 18 (3·3%) of 542 patients in group 2. Using the microbiological reference standard, FujiLAM sensitivity was 60% (95% CI 51–69) and AlereLAM sensitivity was 40% (31–49; p<0·001). Among patients with CD4 counts of less than 200 cells per μL, FujiLAM sensitivity was 69% (57–79) and AlereLAM sensitivity was 52% (40–64; p=0·0218). Among patients with CD4 counts of 200 cells per μL or higher, FujiLAM sensitivity was 47% (34–61) and AlereLAM sensitivity was 24% (14–38; p=0·0116). Using the microbiological reference standard, FujiLAM specificity was 87% (95% CI 85–89) and AlereLAM specificity was 86% (95 CI 84–88; p=0·941). FujiLAM sensitivity varied by lot number from 48% (34–62) to 76% (57–89) and specificity from 77% (72–81) to 98% (93–99). INTERPRETATION: Next-generation, higher sensitivity urine-lipoarabinomannan assays are potentially promising tests that allow rapid tuberculosis diagnosis at the point of care for HIV-positive patients. However, the variability in accuracy between FujiLAM lot numbers needs to be addressed before clinical use. FUNDING: ANRS and Médecins Sans Frontières.