BACKGROUND: A vast amount of scientific research is directed towards the beneficial effects of antioxidants on health. For this reason, several assays have been developed to determine the total antioxidant capacity of blood. METHODS: In this study two procedures based on the use of the green-blue 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) radical (ABTS(*+)) were compared. In the first (commercially available) procedure, ABTS(*+) was generated in the presence of the blood sample. In the second procedure, referred to as the decolorization assay, antioxidants react with preformed ABTS(*+). RESULTS: It was found that the first procedure leads to greater underestimation of the actual antioxidant capacity and is more prone to artifacts than the second procedure. Therefore, only the latter procedure was evaluated in detail and it appeared that (i) plasma is preferred over serum, (ii) the high background produced by albumin can be circumvented by deproteination, (iii) samples can be stored at -80 degrees C for 12 months, and (iv) the assay has high precision. Due to poor linearity, the procedure has to be standardized to allow sample comparison. CONCLUSIONS: The decolorization assay is a reliable and robust assay that can be applied routinely to predict the antioxidant capacity of blood.