Abstract

BACKGROUND:

Tuberculosis (TB) in Africa is increasing because of the human immunodeficiency virus (HIV) epidemic, and in HIV/AIDS patients it presents atypically. Pulmonary tuberculosis (PTB) in Africa is mainly diagnosed clinically, by chest radiograph or by sputum smear for acid fast bacilli (AFB).

METHODS:

We evaluated in 120 HIV-infected patients with chest infection the diagnostic accuracy of AFB smear of sputum and bronchoalveolar lavage (BAL) fluid, sputum Mycobacterium tuberculosis (MTB) culture, real-time PCR and MycoDot serological test, using MTB culture of BAL fluid as gold standard. We correlated PCR cycle threshold values (C(T)) to the culture results. Retrospectively, we evaluated the development of active TB in patients with positive PCR but negative culture.

RESULTS:

Culture of BAL fluid identified 28 patients with PTB. Fifty-six patients could not produce adequate sputum. Sputum AFB smear and the serological test had sensitivities of 66.7% and 0%, respectively. PCR with C(T) 40 was positive in 73 patients, 27 of whom were also TB culture positive (96.4% sensitivity and 52.3% specificity of PCR). PCR with C(T) 32 had sensitivity of 85.7% and specificity of 90.9% to diagnose PTB in BAL. No patients with positive PCR but negative culture developed active TB during 18 months follow-up.

CONCLUSION:

In these HIV-infected patients, AFB smear and serology had very low sensitivities. PCR of BAL with C(T) value 32 had improved specificity to diagnose active PTB. A prospective follow-up study is warranted in TB/HIV endemic settings, applying real time PCR to both sputum and BAL.