Urine lipoarabinomannan (LAM) antigen testing is a rapid TB diagnostic but its utility in TB screening and treatment strategies is unclear. We evaluated the diagnostic yield of urine LAM in advanced HIV-infected adults screened out for TB using routine measures, outcomes in those who tested positive, and the effect of excluding LAM+ participants on trial outcomes.
ACTG A5274 is a randomized clinical trial that demonstrated that four-drug empiric TB therapy did not improve 24-week survival and was associated with an increased incidence of TB compared to isoniazid preventive therapy (IPT) (Figure 1A) in HIV-infected individuals initiating ART with CD4 counts <50 cells/mm3. Participants were screened for TB prior to entry using a symptom screen, locally available diagnostics, and Xpert MTB-RIF when available. Retrospective testing using the Alere urine LAM antigen assay was performed on stored urine obtained prior to ART initiation (baseline); tests were positive if two readers agreed. We determined the proportion of participants with a positive urine LAM test and the effect of excluding these participants on 24-week survival and TB incidence. Kaplan Meier method was used to estimate the primary endpoint (death or unknown vital status) and TB incidence rates at week 24, and the rates were compared by the z-test. Time to confirmed or probable TB (verified by external review) was compared by the log rank test.
Overall, 850 of 1368 candidates were enrolled. Of those screened out, 174 (34%) had suspected TB. Of the 850 enrolled, 53% were male, the median age was 36 years, and the median baseline CD4 count was 18 cells/mm3. Of the 850, 566 (67%) provided baseline urine samples that were tested for LAM antigen (283 in each arm); 28 (5%) were positive [21 (7%) and 7 (2%) in the Empiric and IPT arms, respectively]. Of those positive, 1 participant in each arm died and 5 of 21 and 0 of 7 in the Empiric and IPT arms, respectively, developed TB. After excluding these 28 from the analysis, there were 21 primary endpoints (5%) in each arm (p=0.9). The incidence of TB remained higher (4.6% vs. 2%, p=0.04) and the time to TB remained faster in the Empiric arm (p=0.04; Figure 1B; unchanged in competing risk analysis).
Addition of urine LAM testing is unlikely to yield a positive test among outpatients with advanced HIV who are systematically screened for TB. Use of these tests in similar settings is unlikely to change the lack of an effect of empiric TB treatment on survival.