INTRODUCTION Cytotoxic mechanisms are present in the majority of delayed drug hypersensitivity reactions, but are not used as a diagnostic tool. OBJECTIVES The aim of the study was to compare cytotoxic‑based assays with a proliferation assay and drug patch tests in patients with maculopapular eruptions induced by antiepileptic drugs. Peripheral blood mononuclear cells of 23 patients and 24 controls exposed to the drugs were cultured under defined conditions. A drug‑specific response was assessed by measuring granzyme B (GrB) release with an enzyme‑linked immunospot assay, intracellular expression of granulysin (Grl) in CD3-NKp46+ cells with flow cytometry, perforin concentrations in cell culture supernatants with an enzyme‑linked immunosorbent assay, and using the lymphocyte proliferation test. Patch tests with culprit drugs were done in all patients. Lymphocyte proliferation, GrB release, and Grl expression were significantly higher in patients than in controls, while perforin concentrations were not elevated. The sensitivities were 30.4%, 55%, 39.1%, and 17.4% for proliferation, GrB, Grl, and perforin‑based assays, respectively. A significantly higher rate of positive results was observed when assays were done within 2 years after a drug‑induced reaction. The specificities of all assays remained in the range of 95.8% to 100%. The results of patch tests were positive only in 3 patients (sensitivity, 14.3%) and negative in all controls. In vitro assays based on the detection of Grl, and in particular of GrB, are superior to routine diagnostic tests in patients with hypersensitivity to antiepileptic drugs. They can detect a low‑level response that might be overlooked by standard techniques. In the remission phase, drug‑specific cells are more easily detectable directly in the circulation than in the skin.