Symptom and chest radiographic screening for infectious tuberculosis prior to starting isoniazid preventive therapy: yield and proportion missed at screening.

OBJECTIVE: This analysis describes the prevalence of and risk factors for tuberculosis at screening prior to isoniazid preventive therapy (IPT); the additional yield of tuberculosis using chest radiography versus symptoms alone, and risk factors for tuberculosis missed by screening. DESIGN: Cross-sectional analysis of a trial of community-wide IPT in South African gold mines. METHODS: Participants were screened for tuberculosis prior to starting IPT using symptoms (cough >2 weeks, weight loss, night sweats) and chest radiography. Tuberculosis suspects had sputum collected for mycobacterial investigations. Those with a positive smear or culture with no speciation or culture identified as Mycobacterium tuberculosis were classified as having probable or definite tuberculosis, respectively. Among participants who were dispensed IPT, we defined a 'missed' case of active tuberculosis as one identified within 90 days of the enrolment screen. RESULTS: Between July 2006 and December 2008, among 23,286 participants with complete data, the prevalence of undiagnosed tuberculosis [definite (284) and probable (31)] was high (315/23 286; 1.4%). The addition of chest radiography to symptom screening increased the number of definite tuberculosis cases detected by 2.5-fold (113 to 281 cases). Among 19,609 individuals correctly screened for tuberculosis who started IPT and had more than 90 days of follow-up, only 39 (0.2%) active tuberculosis cases were missed. Risk factors for tuberculosis missed by screening included increasing age [adjusted odds ratio (aOR) 1.66/10 year increase, 95% confidence interval (CI) 1.07-2.56], non-South African, in HIV care (aOR 4.80, 95% CI 1.63-14.1), lower weight (aOR 2.07/10 kg decrease, 95% CI 1.23-3.49) and alcohol use (aOR 2.52, 95% CI 1.31-4.86), which were similar to risk factors for tuberculosis detected by screening. CONCLUSION: Tuberculosis screening prior to IPT detects a substantial burden of tuberculosis and misses very few cases. Chest radiography significantly increased the yield of tuberculosis cases detected.