RÉSUMÉ
BACKGROUND: Interferon-gamma release assays (IGRAs) may be useful in diagnosing latent tuberculous infection (LTBI) in inmates; however, published experience in these settings is limited. OBJECTIVE: To identify variables associated with IGRA positivity among Canadian federal inmates with positive tuberculin skin test (TST) results. DESIGN: On intake, TST-positive (≥10 mm) inmates were offered an IGRA (QuantiFERON(®)-TB Gold), and demographic and historical data were collected. IGRA-positive and -negative inmates were compared using the χ(2) test and multivariable logistic regression; the final model's goodness of fit was assessed using Hosmer-Lemeshow test and area under the receiver operating characteristic curve (AUC). RESULTS: Of 96 TST-positive inmates, 31 (32.3%) were IGRA-positive. Variables associated with positive IGRA were age >45 years (11/20 vs. 20/75, P = 0.016) and previous LTBI treatment (9/20 vs. 13/55, P = 0.032) in univariate analysis, and being from a country with a moderate or high estimated tuberculosis (TB) incidence (OR 3.5, 95%CI 1.3-9.4, P = 0.013) and absence of bacille Calmette-Guérin (BCG) vaccination (OR 3.3, 95%CI 1.2-9.0, P = 0.017) in multivariable analysis. The data fit the model well, classifying the group better than chance alone (AUC 0.67, P = 0.007). CONCLUSION: High discordance with TST, particularly among BCG-vaccinated inmates and those from low TB incidence countries, suggest that IGRA may be useful in Canadian federal penitentiary screening programmes.