The World Health Organization (WHO) symptom screening algorithm for tuberculosis had low sensitivity in identifying TB among South African HIV patients, especially those taking antiretroviral therapy (ART). The findings have implications for starting isoniazid preventive therapy (IPT).
Current symptom algorithms that screen for tuberculosis before starting IPT were developed in patients naive to ART. Researchers in South Africa conducted this study to validate the WHO screening algorithm in patients on ART.
This cross-sectional analysis involved 1429 people, 54% of them on ART, who had symptom screening and a sputum culture in a South African HIV clinic. Clinicians diagnosed culture-positive TB in 126 patients (8.8%, 95% confidence interval [CI] 7.4% to 10.4%).
The WHO symptom screening algorithm had low sensitivity in antiretroviral-naive people (47.6%) and even lower sensitivity in people on ART (23.8%). Specificity was much higher—79.8% in the antiretroviral-naive group and 94.4% in the ART group. (Sensitivity is the ability of a test to single out people who have a certain disease. Specificity is the ability of the test to classify people who do not have the disease as negative.)
The impact of ART on the WHO symptom screen was independent of CD4 count.
Among people in whom the WHO algorithm did not indicate TB, 8.9% (95% CI 7.4% to 10.8%) of the antiretroviral-naive group and 4.4% (95% CI 3.7% to 5.2%) of the treated group actually had culture-confirmed TB.
Adding body mass index to the WHO symptom screen significantly improved its ability to discriminate between presence and absence of TB in both antiretroviral-naive and treated people. Adding CD4 count and ART duration further improved the discriminatory power of the WHO algorithm.
The researchers conclude that “the WHO symptom screen has poor sensitivity, especially among patients on ART, in a clinic where regular tuberculosis screening is practiced.” As a result, they caution, “a significant proportion of individuals with tuberculosis would inadvertently be placed on isoniazid monotherapy despite high negative predictive values.”
“Until more sensitive methods of ruling out tuberculosis are established,” the authors advise, “it would be prudent to do a sputum culture prior to IPT where this is feasible.”
Source: Molebogeng X. Rangaka, Robert J. Wilkinson, Judith R. Glynn, Andrew Boulle, Gilles van Cutsem, Rene Goliath, Shaheed Mathee, Gary Maartens. Effect of antiretroviral therapy on the diagnostic accuracy of symptom screening for intensified tuberculosis case finding in a South African HIV clinic. Clinical Infectious Diseases. 2012; 55: 1698-1706.
(Downloading the complete article requires a subscription to Clinical Infectious Diseases or an online payment; the abstract is free.)
Source: IAS