Targeted screening for tuberculosis infection in priority populations in Guangzhou, China: IGRA positivity, associated factors, and risk interactions

Community tuberculosis screening programmes in China prioritise several population groups, but interferon-gamma release assay (IGRA) positivity and its associated factors may differ substantially across groups. We assessed IGRA positivity among residents screened in 11 programme communities in Guangzhou and examined patterns in the school population, older adults, and people with self-reported diabetes. We analysed cross-sectional screening data from 12,299 participants using site-clustered GEE for multivariable and stratified analyses, and evaluated three hypothesis-driven exploratory interactions on the multiplicative and OR-based additive scales with simulation-based confidence intervals. Sensitivity analyses excluded participants with prior tuberculosis and applied Firth penalised logistic regression to the sparse diabetes × tuberculosis-contact interaction. Overall IGRA positivity was 14.48% (1,781/12,299). Positivity was 5.32% (254/4,771) among students, 11.83% (117/989) among school staff, 23.44% (1,019/4,348) among adults aged ≥ 60 years, and 32.96% (614/1,863) among people with self-reported diabetes. In the site-clustered model, male sex, older age, reported tuberculosis contact, prior tuberculosis, self-reported diabetes, other comorbidities, and suspected tuberculosis symptoms were associated with IGRA positivity. Positive interactions were observed for diabetes × male sex (RERI 2.63, 95% CI 1.82–3.56), diabetes × tuberculosis contact (RERI 20.75, 95% CI 4.14–70.72), and male sex × older age (RERI 1.25, 95% CI 0.82–1.73). For diabetes × tuberculosis contact, the multiplicative interaction remained significant in the non-clustered Firth analysis (OR 3.41, 95% CI 1.27–11.29; P = 0.014). IGRA positivity differed substantially across programme groups. The findings support prioritising people with diabetes and recent tuberculosis contact, older adults with additional risk factors, and exposed students and school staff for clinical evaluation. The interaction findings were exploratory and should not be interpreted as causal effects.