The microtiter plate technique reported by Baker and colleagues for the glutathione reductase-DTNB recycling assay of total glutathione (GSx) and glutathione disulfide (GSSG) has been modified according to Anderson's recommendations, in order to improve the reliability and accuracy of this miniaturized method for the measurement of glutathione status in cultured/isolated cells. Dilute HCl (10 mmol/L) has been used to lyse cells, before protein removal by centrifugation in the presence of 1.3% sulfosalicylic acid. The final DTNB, GSSG-reductase and NADPH concentrations in the reaction mixture have been increased to 0.7 mmol/L, 1.2 IU/ml and 0.24 mmol/L, respectively. The procedure specificity has been tested by spiking and dilution assays, showing that about 90% of the expected GSx amounts could actually be recovered, while no changes of GSSG concentrations were caused in the cells. Accuracy has been assessed by analysis of within-series precision as well as of intra- and interassay reproducibility, showing coefficient variation of <10%. Glutathione changes measured either in control rat hepatocytes or in primary cultures treated with paracetamol or menadione were in good agreement with well-known literature data. These data suggest that the experimental conditions reported in this paper are suitable for the analysis of total glutathione and glutathione disulfide concentrations in cultured/isolated cells.