Community-wide annual tuberculosis screening over four years: Findings from the HPTN 071 (PopART) trial

Background

The World Health Organization suggest that systematic tuberculosis (TB) screening may be conducted in high prevalence settings (>0.5%), though supporting evidence is limited.

Methods

Between January-2014 to December-2017, the HPTN 071 (PopART) cluster-randomized trial implemented universal HIV and TB testing across 21 communities in Zambia and South Africa (SA), with TB prevalence of 0.5% and 1.6%, respectively. Trained community health workers visited households annually to offer HIV testing and TB symptom screening, with sputum collection from individuals who screened positive. Diagnostic testing used Xpert-MTB/RIF or smear microscopy, and linkage to treatment was facilitated.

We analysed TB screening and diagnosis data across three rounds (R1–R3) in Zambia and R3 in SA, where complete data were available. We examined factors associated with reporting TB symptoms and being diagnosed with TB.

Results

The yield of newly diagnosed TB (per 100,000 persons) increased across rounds in Zambia [R1 = 81, R2 = 93, R3 = 110; p-value (trend) = 0.003] and was higher in R3 SA (380). In R3, TB yield was higher in men (Zambia: 146; SA: 543) than women (Zambia: 76; SA: 257), and among newly diagnosed HIV-positive individuals (Zambia: 541; SA: 789) compared to HIV-negative individuals (Zambia: 48; SA: 170) and self-reported HIV-positive individuals on ART (Zambia: 105; SA: 192). In Zambia R3, participants screened twice before had 38% lower odds of being diagnosed with TB compared to those screened first time [Adjusted odds ratio = 0.62, 95% CI (0.43, 0.90)].

Conclusion

The PopART intervention identified undiagnosed TB through systematic TB symptom screening, particularly in men and newly diagnosed HIV-positive individuals and individuals who had not previously been screened. Yield was relatively low compared to estimated TB prevalence.

Trial registration

ClinicalTrials.gov NCT01900977