Empiric tuberculosis treatment in hospitalised adults with advanced HIV disease in Africa: applying the therapeutic threshold to clinical practice.

Hospitalised adults with advanced HIV disease in sub-Saharan Africa experience high mortality, with tuberculosis, often disseminated and undiagnosed, being a leading cause. Despite this, initiation of antituberculous therapy is frequently delayed pending diagnostic confirmation, which may be unavailable in this population. Evidence from recent trials and cohort studies suggests that even short delays in antituberculous therapy are associated with substantial increases in mortality, whereas empiric therapy might improve survival in people at high risk. Applying the therapeutic threshold framework, the high pre-test probability of tuberculosis in severely ill inpatients with advanced HIV disease often exceeds the threshold at which treatment benefits outweigh risks, even in the absence of confirmatory testing. Although concerns regarding toxicity, drug interactions, and overtreatment are valid, short-term empiric antituberculous therapy appears safe and these risks might be outweighed by the consequences of untreated disease. We argue for a paradigm shift towards earlier empiric antituberculous therapy, with parallel diagnostic evaluation and structured reassessment in selected patients, and for adequately powered randomised controlled trials of empiric therapy powered for mortality.