Screening for Asymptomatic Tuberculosis among Adults with Household Exposure to a Patient with Pulmonary Tuberculosis

More than half of tuberculosis (TB) detected by community prevalence surveys is classified as asymptomatic. We evaluated yield of symptom and chest radiograph (CXR) screening of TB-exposed household contacts (HHC) in South Africa. Adult volunteers (≥18 years) with household exposure to pulmonary TB patients were enrolled at three sites. Systematic screening of TB symptoms (any duration), CXR (any abnormality), and sputum microscopy, Xpert Ultra, and liquid culture were performed. Serum C-reactive protein (CRP) was measured by multiplex bead array. Prevalent TB was microbiologically-confirmed (Xpert Ultra or culture). Symptomatic and asymptomatic TB were defined as prevalent TB with and without reported symptoms, respectively. Between March 2021 - December 2022, 979 HHC were enrolled; 185 (18.9%) living with HIV and 187 (19.1%) with previous TB. Prevalent TB occurred in 51 (5.2%) and was asymptomatic in 42/51 (82.4%). Only 13/42 (31.0%) asymptomatic TB cases were smear-positive [8/13 (61.5%) graded scanty or 1+]. CRP did not discriminate healthy HHC from those with asymptomatic TB (AUC 0.60; 95%CI 0.47-0.73). An abnormal CXR was observed in 23/41 asymptomatic (sensitivity 56.1%, 95%CI 41.0-70.1%) versus 8/9 symptomatic (sensitivity 88.9%, 95%CI 56.5-98.0%) TB cases. Sensitivity of CXR in combination with symptom screening was 64.0% (32/50, 95%CI 50.1-75.9%) for all prevalent TB. More than 80% of confirmed TB cases among HHC were asymptomatic. CXR screening missed more than 40% of these asymptomatic cases. Community prevalence surveys reliant on symptom- and CXR-based approaches may significantly underestimate the prevalence of asymptomatic TB in endemic countries. Supported by RePORT South Africa through funding from the U.S. National Institutes of Health, CRDF Global, and the South African Medical Research Council. Evidence before this study: World Health Organisation (WHO) guidelines for systematic tuberculosis (TB) screening recommend symptom screening and chest radiography (CXR), based on a Cochrane meta-analysis reporting 70.6% sensitivity (any TB symptom) and 94.7% sensitivity (any CXR abnormality) for bacteriologically-confirmed pulmonary TB. National TB prevalence surveys rely on a positive symptom screen or abnormal CXR to trigger diagnostic sputum testing. This approach to community screening would, by definition, miss asymptomatic TB cases without CXR evidence of disease. We reviewed the reference list of the aforementioned meta-analysis for active case-finding studies of adolescents and adults aged 15 years and older in community and contact-tracing settings. We performed forward citation-tracking and searched reference lists, including studies published in English between Jan 1, 1980, and November 1, 2024. We excluded studies that included children <15 years; or that exclusively enrolled people with additional risk factors (HIV; diabetes; latent TB infection; prior TB). We found 28 studies that performed universal sputum testing for bacteriologically-confirmed pulmonary TB and reported 51.8% (95%CI 49.9-53.7%; I 2 = 89.2%) pooled sensitivity for symptom screening (any symptom; 24 studies, 2,969 TB cases) and 62.4% (95%CI 59.3-65.3%; I 2 = 88.3%) pooled sensitivity for CXR (any abnormality; 10 studies, 1,123 TB cases). Only four studies (145 TB cases) reported accuracy of symptom screening in parallel with chest radiography (pooled sensitivity 67.3%, 95%CI 57.3-75.9%; I 2 = 87.1%), but these studies did not disaggregate symptomatic and asymptomatic disease. Added value of this study: We performed systematic screening using universal sputum microbiological testing of 978 household contacts of pulmonary TB patients in three South African communities and compared symptom (any duration) and CXR (any abnormality) screening approaches against a microbiological reference standard. We detected confirmed pulmonary TB in 5.2% of household contacts, and 82.4% of these TB cases reported no TB symptoms. Asymptomatic TB in household contacts was pauci-bacillary and associated with low serum CRP levels that were indistinguishable from healthy controls, but distinct from symptomatic TB in a comparator group of clinic attendees. Sensitivity of CXR screening for asymptomatic TB was only 56.1%; sensitivity of combined symptom and CXR screening for all TB was marginally higher at 64.0%.Implications of all the available evidence: Our findings from household contacts suggest that symptom- and CXR-based approaches are inadequate for community TB screening in South Africa and do not meet the WHO Target Product Profile for a TB screening test (minimum 90% sensitivity; 70% specificity). National TB Prevalence Surveys that omit universal sputum microbiological testing may significantly underestimate the prevalence of asymptomatic TB in high-burden countries.