Point-of-care urine LAM testing to guide tuberculosis treatment among severely ill in-patients with HIV in real-world practice: a multi-center stepped wedge cluster-randomized trial from Ghana.

Abstract

The lateral flow urine lipoarabinomannan assay, Determine TB-LAM (Determine LAM), offers the potential for timely tuberculosis (TB) treatment among people living with HIV (PWH). In this cluster-randomized trial, Determine LAM was made available with staff training with performance feedback at three hospitals in Ghana. Newly admitted PWH with a positive WHO four-symptom screen for TB, severe illness, or advanced HIV were enrolled. The primary outcome was days from enrollment to TB treatment initiation. We also reported the proportion of patients with a TB diagnosis, initiating TB treatment, all-cause mortality, and Determine LAM uptake at 8 weeks. We enrolled 422 patients including 174 (41.2%) in the intervention group. The median CD4 count was 87 cells/mm3 (IQR 25-205) and 138 patients (32.7%) were on antiretroviral therapy. More patients were diagnosed with TB in the intervention group compared with the control group, 59 (34.1%; 95%CI27.1-41.7) vs 46 (18.7%; 95%CI14.0-24.1), p <0.001. Time to TB treatment remained constant at a median of 3 days (IQR 1-8), but patients were more likely to initiate TB treatment, aHR 2.19 (95%CI1.60-3.00) during the intervention. Of patients with a Determine LAM test available, 41 (25.3%) tested positive. Of those, 19 (46.3%) initiated TB treatment. Overall, 118 patients had died (28.2%; 95%CI24.0-33.0) at 8 weeks follow-up. The Determine LAM intervention in real-world practice increased TB diagnosis and the probability of TB treatment but did not reduce time to treatment initiation. Despite high uptake, only half of the LAM-positive patients initiated TB treatment.