Effect of Variability on Interferon-Gamma Release Assay Performance: A Quantitative Analysis.

Abstract

Interferon-gamma release assays (IGRAs) are widely used to identify latent tuberculosis infection (LTBI); however, inherent test variability affects their diagnostic interpretation. We evaluated false-positive and false-negative rates, as well as conversion and reversion rates across CVs of 20%, 40%, 60%, 80%, and 100%, using statistical modeling. At a diagnostic cutoff of 0.35 IU/mL, false-negative rates increased from 1.61% to 33.41% with increasing CVs, whereas false-positive rates ranged from 0.00% to 15.87% within the 0.20-0.70 IU/mL borderline range. Expanding the borderline to 0.20-1.00 IU/mL reduced false-positive rates to a maximum of 3.16%, without affecting false-negative rates. Within the 0.20-0.70 IU/mL borderline zone, correct reversion and false conversion rates at 0.20 and 0.35 IU/mL ranged from 0.01% to 25.00% and 0.00% to 24.20%, respectively. At 0.35, 0.70, and 1.00 IU/mL, correct conversion and false reversion rates ranged from 0.05% to 24.20% and 0.00% to 25.00%, respectively. These results highlight the importance of adopting borderline zones in IGRA interpretation to reduce misclassification, although variability from manufacturing, pre-analytical processing, and analytical procedures remains a significant challenge. Reducing such variability through improved production consistency, standardized sample handling, and automated analysis platforms is essential to enhance the diagnostic reliability of IGRAs for LTBI.