Background. Interferon-γ release assays (IGRAs) have been widely used for the diagnosis of latent and active tuberculosis in adults, but their role in diagnosing latent tuberculosis infection (LTBI) in children younger than 5 years remains unclear. Objective. To evaluate the diagnostic performance of IGRAs for LTBI in children younger than 5 years. Methods. We searched the PubMed, EMBASE, Cochrane Library, and Web of Knowledge databases. According to inclusion criteria, the diagnostic studies of IGRAs for LTBI in children younger than 5 years were included. The study quality was assessed using the 14-item Quality Assessment of Diagnostic Accuracy Studies instrument. The Meta-Analyst software was used to analyze the data. Results. Seven studies (627 patients) were included. Meta-analysis showed that compared with tuberculin skin test (TST), the pooled sensitivity (Sen), specificity (Spe), positive likelihood ratio (PLR), negative likelihood ratio (NLR), positive predictive value (PPV), negative predictive value (NPV), accuracy and diagnosis odds ratio (DOR), and summary receiver operating characteristic (SROC) curve of QuantiFERON-TB (values in parentheses are 95% confidence intervals) were 0.841 (0.733-0.911), 0.895 (0.861-0.922), 8.046 (5.953-10.876), 0.192 (0.113-0.326), 0.622 (0.527-0.709), 0.970 (0.946-0.983), 0.891 (0.861-0.916), 50.718 (24.201-106.287), and 0.809 respectively. The pooled Sen, Spe, PLR, NLR, PPV, NPV, accuracy, DOR, and SROC curve of T-SPOT.TB test were 0.931 (0.760-0.983), 0.767 (0.684-0.833), 3.947 (2.854-5.459), 0.095 (0.026-0.354), 0.523 (0.398-0.645), 0.977 (0.914-0.994), 0.806 (0.737-0.860), 65.270 (13.463-316.434), and 0.855, respectively. Conclusion. This meta-analysis showed that the sensitivity and specificity of IGRAs were slightly higher than TST and can be used as supporting tools to detect LTBI in children younger than 5 years. The methodological quality of diagnostic test needs to be improved. [ABSTRACT FROM PUBLISHER]