Clinical diagnosis of TB: lessons on misdiagnosis and overdiagnosis.

Abstract

Clinically diagnosed TB patients (n = 335) at two facilities in Lusaka, Zambia were re-evaluated within two weeks of diagnosis. This re-evaluation included sputum Xpert Ultra testing and expert reader interpretation of the chest x-rays (CXRs) used for initial diagnosis. Repeat Xpert Ultra detected TB in just 2.6% (n=6). Of the remaining patients (n=222), expert CXR re-interpretation classified 18.0% as normal; 36.0% as abnormal, consistent with TB; and 46.0% as abnormal, not consistent with TB. These findings suggest that clinical TB is frequently over diagnosed in those without detectable CXR abnormalities and misdiagnosed in those with abnormal CXRs: these abnormalities are likely due to other respiratory conditions. Such misdiagnosis leads to unnecessary treatment, failure to treat the true underlying condition and incorrect estimates of TB burden.