RATIONALE: Interferon gamma (IFN-g) enzyme-linked immunospot (ELISpot) assay has been introduced for drug hypersensitivity diagnosis, however, the cut-off value is not yet optimized. The aim is to determine the clinical cut-off value of IFN-g ELISpot assay in patients with a history of drug hypersensitivity based on drug challenge test and/or skin test results. METHODS: The frequencies of IFN-g releasing cells after stimulating peripheral blood mononuclear cells (PBMC) with 72 suspected culprit drug(s) in 54 patients with a history of drug hypersensitivity were comparatively analyzed with the results of drug challenge test and/or skin test. The optimal cut-off value was calculated based on the receiver operating characteristic (ROC) curve results. RESULTS: About 40.7% (22/54) of patients in this study have a history of severe cutaneous adverse reactions (SCARs) including 8 drug reaction with eosinophilia and systemic symptoms, 5 Stevens-Johnson syndrome/ toxic epidermal necrolysis, and 9 acute generalized exanthematous pustulosis. The average frequencies of drug-induced IFN-g releasing cells in confirmed drug hypersensitivity group were significantly higher than those in drug tolerant group (70.2623.4 and 13.7610.8 cells/10 6 PBMC), respectively; p value50.04). According to ROC curve analysis, the cut-off value of 22 cells/10 6 PBMC yielded 65.2% sensitivity, 95.9% specificity, and excellent positive likelihood ratio (15.9). The sensitivity and specificity increased to 77.8% and 100.0%, respectively, in drug-induced SCAR subjects. CONCLUSIONS: Our study confirms that IFN-g ELISpot assay has good clinical diagnostic values to identify the culprit drug, especially in SCARs. The suspected culprit drugs with positive ELISpot result should be avoided from further use due to high specificity of the test.79 Hypersensitivity reactions to antituberculosis drugs confirmed by interferon gamma enzymelinked Immunospot assay