“…In MTB endemic settings, anti-tubercular drugs may be initiated however this approach may risk adverse drug reactions and interactions, hepatotoxicity and miss other pathology (Mendelson 2007). Indeed, a study from India using EUS FNA to assess abdominal lymphadenopathy in HIV-uninfected patients, confirmed that almost half of the patients in this high endemic region were found to have non-tubercular pathology (Dhir et al 2011) Similarly, our study confirmed that 31% of HIV infected patients with enlarged nodes, even in a high endemic area such as South Africa, had reactive lymphadenopathy. This implies that empirical ATT, prescribed solely on the basis of the presence of enlarged lymph nodes, may be associated with considerable treatment-related risks and should consequently be avoided.…”