Prevalence of subclinical pulmonary tuberculosis and its association with HIV in household contacts of index tuberculosis patients in two South African provinces: a secondary, cross-sectional analysis of a cluster-randomised trial

Background: People with subclinical tuberculosis (TB) have microbiological evidence of disease caused by Mycobacterium tuberculosis, but either do not have, or do not report TB symptoms. The relationship between human immunodeficiency virus (HIV) and subclinical TB is not yet well understood. We estimated the prevalence of subclinical pulmonary TB in household contacts of index TB patients in two South African provinces, and how this differed by HIV status. Methods: This was a cross-sectional, secondary analysis of baseline data from the intervention arm of a household cluster randomised trial. Prevalence of subclinical TB was measured as the number of household contacts aged ≥5 years who had positive sputum TB microscopy, culture, or nucleic acid amplification test (Xpert MTB/Rif or Xpert Ultra) results on a single sputum specimen and who did not report current cough, fever, weight loss or night sweats on direct questioning. Regression analysis was used to calculate odds ratios (OR) and 95% confidence intervals (CI) for the association between HIV status and subclinical TB, adjusting for province, sex, and age in household contacts, and HIV status in index patients. Results: Amongst household contacts, microbiologically-confirmed prevalent subclinical TB was over twice as common as symptomatic TB disease (48/2077, 2.3%, 95% CI 1.7–3.1% compared to 20/2077, 1.0%, 95% CI 0.6–1.5%). Subclinical TB prevalence was higher in people living with HIV (15/377, 4.0%, 95% CI 2.2–6.5%) compared to those who were HIV-negative (33/1696, 1.9%, 95% CI 1.3–2.7%; p=0.018). In regression analysis, living with HIV (377/2077, 18.2%) was associated with a two-fold increase in prevalent subclinical TB with 95% confidence intervals consistent with no association through to a four-fold increase (adjusted OR 2.00, 95% CI 0.99–4.01, p=0.052). Living with HIV was associated with a five-fold increase in prevalent symptomatic TB (adjusted OR 5.05, 95% CI 2.22–11.59, p<0.001). Conclusions: Most (70.6%) pulmonary TB diagnosed in household contacts in this setting was subclinical. Living with HIV was likely associated with prevalent subclinical TB, and was associated with prevalent symptomatic TB. Universal sputum testing with sensitive assays improves early TB diagnosis in subclinical household contacts