Abstract
BACKGROUND: HIV-exposed uninfected (HEU) infants have increased risk of tuberculosis (TB). Testing for Mycobacterium tuberculosis (Mtb) infection is limited by reduced Quantiferon (QFT) sensitivity in infants and tuberculin skin test (TST) cross-reactivity with Bacillus Calmette-Guérin vaccine. Our objective is to assess if non-IFNγ cytokine responses to Mtb-specific antigens have improved sensitivity in detecting Mtb infection in HEU infants compared with QFT. METHODS: HEU infants were enrolled in a randomized clinical trial of isoniazid preventive therapy (IPT) to prevent Mtb infection in Kenya (N = 300) and assessed at 12 months postrandomization (14 months of age) by TST and QFT-Plus. Non-IFNγ cytokine secretion (IL2, TNF, IP10, N = 229) in QFT-Plus supernatants was measured using Luminex assay. Logistic regression was used to assess the effect of IPT on Mtb infection outcomes in HEU infants. RESULTS: Three of 251 (1.2%) infants were QFT-Plus positive. Non-IFNγ Mtb antigen-specific responses were detected in 12 additional infants (12/229, 5.2%), all TST negative. IPT was not associated with Mtb infection defined as any Mtb antigen-specific cytokine response (odds ratio = 0.7, P = 0.54). Mtb antigen-specific IL2/IP10 responses had fair correlation (τ = 0.25). Otherwise, non-IFNγ cytokine responses had minimal correlation with QFT-Plus and no correlation with TST size. CONCLUSIONS: We detected non-IFNg Mtb antigen-specific T-cell responses in 14-month HEU infants. Non-IFNg cytokines may be more sensitive than IFNg in detecting infant Mtb infection. IPT during the first year of life was not associated with Mtb infection measured by IFNg, IL2, IP10 and TNF Mtb-specific responses.