Background: The diagnosis of TB in HIV-infected patients is challenging. Both urinary lipoarabinomannan (Alere TB™ LAM) and GeneXpert-MTB/RIF (Xpert) are useful for the diagnosis of TB. However, how to optimally integrate Xpert and LAM into clinical practice algorithms remain unclear.
Methods: We performed a post-hoc analysis of 561 HIV-infected sputum expectorating patients (median CD4 count of 130 cells/ml) from a previously published randomised controlled trial evaluating LAM in hospitalized HIV-infected patients with suspected TB. We evaluated 5 different diagnostic strategies using sputum culture as a reference standard [Xpert alone, LAM alone, sequential Xpert followed by LAM and vice versa (LAM in Xpert-ve; Xpert in LAM-ve) or performing both tests concurrently (LAM + Xpert)]. A cost-consequence analysis was performed.
Results: Strategy-specific sensitivity and specificity, using culture as a reference, was similar with the Xpert-only and sequential/concurrent strategies. However, when using any positive TB-specific test as a reference, the incremental yield of LAM over Xpert was 29.6% (45/152) and Xpert over LAM 75% (84/112). The incremental yield of LAM increased with decreasing CD4 count. The cost per TB case were similar for the sequential and concurrent strategies ($1617 to $1626).
Conclusion: In sputum-expectorating hospitalised patients with advanced HIV and access to both tests, concurrent testing with Xpert and LAM may be the best strategy for diagnosing TB. These data inform clinical practice in TB and HIV endemic settings.