WHO TB symptom screen performs poorly in South Africans on ART
A symptom-based tuberculosis screening questionnaire recommended by the World Health Organization (WHO) for people with HIV performed poorly in detecting TB and excluding pulmonary TB in South Africans taking antiretroviral therapy (ART). The symptom screen performed adequately in HIV-positive people not taking antiretrovirals.
Because TB remains the leading cause of HIV morbidity and mortality in sub-Saharan Africa, clinicians without sophisticated diagnostic tests need reliable tools to predict who has TB and who does not. To help identify TB patients on the basis of symptoms, the WHO recommends a symptom-screening questionnaire. This cross-sectional study at two HIV clinics in South Africa aimed to determine the reliability of the WHO screening tool in HIV-positive people on or off ART.
The study involved 522 people taking ART and 215 not taking ART. Everyone completed the screening questionnaire, had microbiologic sputum testing, and had a chest x-ray. The researchers defined a positive TB screen on the questionnaire as cough, fever, night sweats, or weight loss. They defined pulmonary TB as a sputum smear positive for acid-fast bacilli or growth of Mycobacterium tuberculosis.
Sputum smear diagnosed pulmonary TB in 31 of 522 people (5.9%) on ART and 34 of 215 (15.8%) not taking antiretrovirals. Answers on the WHO screening questionnaire missed 15 of 31 cases of pulmonary TB (48.4%) in people on ART but only 3 of 34 cases (8.8%) in people not on ART.
In people taking antiretrovirals, post-test probability of a pulmonary TB diagnosis was 6.8% in those screening positive on the questionnaire and 5.2% in those screening negative. In people not taking antiretrovirals, post-test probability of pulmonary TB was 20.3% in those screening positive on the questionnaire and 4.8% in those screening negative.
Among people diagnosed with pulmonary TB by sputum testing, patients on ART were 96% less likely to screen positive on the questionnaire than were patients not on ART (adjusted odds ratio 0.04, 95% confidence interval 0.01 to 0.39).
In the two study groups—those on ART and those not on ART—TB screening was more sensitive if it included both the questionnaire and chest x-ray.
“For case detection and exclusion of pulmonary TB,” the researchers conclude, “the WHO-recommended questionnaire performed adequately among [HIV-positive people] not on ART, and poorly among those on ART.” They call for further research “to identify feasible and effective TB screening strategies for [people with HIV] in resource-limited settings.”
Source: Faiz Ahmad Khan, Sabine Verkuijl, Andrew Parrish, Fadzai Chikwava, Raphael Ntumy, Wafaa El-Sadr, Andrea A. Howard. Performance of symptom-based tuberculosis screening among people living with HIV: not as great as hoped. AIDS. 2014; 28: 1463-1472.
For the study abstract
(Downloading the complete article requires a subscription to AIDS or an online payment; the abstract is free.)
Source: IAS
