Method comparison of QIAGEN QuantiFERON-TB gold plus ELISA and LIAISON® QuantiFERON-TB gold plus for latent tuberculosis infection detection in pregnant ugandan women.

BACKGROUND

The DiaSorin LIAISON® QuantiFERON-TB Gold Plus test offers automation advantages over the standard QIAGEN ELISA; however, data comparing the two methods in pregnant women are lacking. We assessed the categorical and quantitative agreement between these interferon-gamma release assay platforms for latent tuberculosis infection (LTBI) detection in a cohort of pregnant women in Uganda.

METHODS

In this cross-sectional diagnostic comparison, 261 pregnant women were enrolled at Kawempe National Referral Hospital in Kampala, Uganda (September-December 2020). Plasma samples were initially tested with QIAGEN ELISA (QFT-ELISA) in 2021, stored at -80°C, and retested in 2025 with both LIAISON® chemiluminescent immunoassay (QFT-CLIA) and QFT-ELISA. Agreement was assessed using Cohen's kappa, Bland-Altman analysis, and Deming regression analyses. Stability over four years was evaluated using weighted kappa and transition analyses. Subgroup analyses examined agreement by HIV status and trimester

RESULTS

Of the 261 enrolled participants, 231 samples (88.5 %) remained viable after four years storage. Overall agreement between QFT-CLIA and QFT-ELISA in 2025 was 90.5 % (Cohen's Kappa 0.82; 95 % CI: 0.75-0.89). Bland-Altman analysis showed minimal systematic bias (mean difference for TB1-Nil: +0.046 IU/mL, p = 0.608, TB2 Nil: +0.0776 IU/mL, p = 0.3175), and Deming regression confirmed linearity (slope, 0.94-0.96). Among the 187 samples that remained QFT-stable between 2021 and 2025, the agreement improved to 98.4 % (kappa 0.97; 95 % CI: 0.94-1.00), with category-specific agreement of 93.6 %, 97.5 %, and 100 % for positive, negative, and indeterminate results, respectively. Four-year stability varied according to baseline classification: 57.1 % for positive samples, 97.5 % for negative samples, and 75.0 % for indeterminate samples. Agreement was similar between HIV-positive (84.6 %, Kappa 0.76) and HIV-negative participants (90.8 %, Kappa 0.82; p = 0.357).

CONCLUSIONS

The LIAISON® platform demonstrated excellent agreement with the QIAGEN ELISA for LTBI detection during pregnancy, particularly when storage effects were minimized. Long-term cryopreservation disproportionately affects positive samples. The automation of the LIAISON® platform supports high-throughput implementation in antenatal care settings; however, fresh sample testing should be prioritized for clinical decision-making.