Should urine-LAM tests be used in TB symptomatic HIV-positive patients when no CD4 count is available? A prospective observational cohort study from Malawi

BACKGROUND: Current eligibility criteria for urine lateral-flow lipoarabinomannan assay (LF-LAM) in ambulatory, HIV-positive patients rely on the CD4 count. We investigated the diagnostic yield of LF-LAM and the 6-month mortality in ambulatory, TB symptomatic, HIV-positive patients regardless of their CD4 count. METHODS: We conducted a prospective, observational study that included all ambulatory, >/=15-year-old, TB symptomatic (cough, weight loss, fever, or night sweats) HIV-positive patients presenting at 4 health facilities in Malawi. Patients received a clinical examination and were requested urine LF-LAM, sputum microscopy, and Xpert MTB/RIF. TB was defined as bacteriologically confirmed if Xpert was positive. RESULTS: Of 485 patients included, 171 (35.3%) had a CD4 /= 200 (12.8% LAM grades 2-4). Xpert was positive in 14.1% (44/312). Among Xpert-positive patients, LAM positivity was 56.7% (CD4 /= 200), P = 0.393. Of the patients without an Xpert result, 13.4% (23/172) were LAM positive (ie, potentially missed patients). Overall, mortality was 9.2% (44/478). More pronounced LAM-positive patients had higher mortality than LAM-negative (grades 2-4: 36.0%; grade 1: 9.1%; negative: 7.4%; P < 0.001). LAM-positive patients with CD4 <200 cells/microL had higher risk of mortality than LAM negatives (adjusted hazard ratio: 3.2, 95% confidence interval: 1.4 to 7.2, P = 0.006), particularly those with LAM grades 2-4 (adjusted hazard ratio: 4.9, 95% confidence interval: 1.8 to 13.3, P = 0.002). CONCLUSIONS: Urine-LAM testing can be useful for TB diagnosis in HIV-positive TB-symptomatic patients with no CD4 cell count. LAM grade can identify patients at higher risk of death in this situation.