Factors associated with interferon-gamma release assay positivity in patients with psoriasis: a cross-sectional study.

BACKGROUND: The interferon-gamma release assay (IGRA) is recommended for latent tuberculosis infection (LTBI) screening prior to biologic therapy in psoriasis. Yet factors influencing IGRA positivity in this population remain insufficiently characterized. OBJECTIVES: This study aimed to identify clinical factors associated with IGRA positivity among patients with psoriasis and to develop an internally validated model to summarize these associations in routine dermatological care. METHODS: We retrospectively analysed 1,126 patients with psoriasis without a documented history of prior tuberculosis (285 IGRA-positive and 841 IGRA-negative) at a single dermatology centre. Variables associated with IGRA positivity in univariable analysis were entered into a multivariable logistic regression model, which was internally validated using bootstrap resampling. RESULTS: The overall IGRA positivity rate was 25.3%. IGRA positivity increased markedly with age, from 8.2% in patients aged ≤ 30 years to 45.6% in those aged > 60 years, with an optimal ROC-derived cutoff of 49 years (sensitivity, 0.561; specificity, 0.762). In multivariable analysis, older age (OR 1.048, 95% CI 1.037-1.059, p < 0.001), pulmonary nodules (OR 1.608, 95% CI 1.094-2.365, p = 0.016), and higher eosinophil percentage (OR 1.109, 95% CI 1.033-1.190, p = 0.004) were independently associated with increased odds of IGRA positivity, whereas prior biologic exposure was associated with reduced odds (OR 0.705, 95% CI 0.512-0.969, p = 0.031). The eight-variable multivariable model demonstrated acceptable discrimination (AUC 0.726; optimism-corrected AUC 0.717, 95% CI 0.690-0.763) and good calibration (Brier score 0.167). Decision curve analysis suggested net clinical benefit across a threshold range of approximately 5%-50%. CONCLUSIONS: Older age, pulmonary nodules, and higher eosinophil percentage were independently associated with IGRA positivity, while prior biologic exposure showed an inverse association likely related to selection bias. The model showed acceptable internal performance and may provide supplementary context, but should not replace standard pre-biologic tuberculosis assessment.