Tuberculosis case finding in HIV-positive adults receiving care at health centres in Ethiopia

Rapid, accurate diagnosis and management of TB in HIV-positive individuals is critical for the success of antiretroviral treatment programmes in primary health care in Sub-Saharan Africa. These studies aim to develop and evaluate methods for TB case finding in adults receiving care at Ethiopian health centres. The prevalence of active TB in a prospective cohort of 812 HIV-positive adults was determined; the diagnostic performances of sputum Xpert MTB/RIF assay and urinary lipoarabinomannan (LAM) antigen detection were evaluated; a clinical algorithm for categorization of TB risk was developed; and TB treatment outcomes in patients identified through active and passive case finding were compared. Active TB was bacteriologically confirmed in 137/812 (16.9%) participants. The Xpert MTB/RIF assay detected two-thirds of culture-positive TB cases, improving case detection rate by 47.4% using smear microscopy as a base case. Urine LAM identified 25.8% of TB cases, but high rates of false-positive reactions were observed. A combination of sputum smear microscopy and urine LAM yielded an overall sensitivity of 66.7% in a subset with CD4 count ≤100 cells/mm3. A clinical scoring algorithm was constructed based on characteristics independently associated with bacteriologically confirmed TB (cough, Karnofsky performance score ≤80, mid-upper arm circumference <20 cm, hemoglobin <10 gm/dL, and peripheral lymphadenopathy) in persons with positive WHO TB symptom screen results. This algorithm could categorize patients into 3 risk groups based on their likelihood of having TB. This strategy would reduce the proportion of individuals in need of further TB investigations by 255/569 (45%). TB/HIV-co-infected patient diagnosed through active case finding had less severe clinical characteristics but treatment outcome remained the same with the cases identified through passive case finding. In this setting, the burden of TB is high, HIV-positive individuals may be categorized using the 2 algorithms, and Xpert MTB/RIF assay is recommended for further TB investigation.