The widely used TBA assay for lipid peroxidation was modified to minimize artefactual oxidative degradation of lipids during the assay. Formation of the TBA-MDA condensation product was studied with and without exclusion of oxygen, and the concentration effect of BHT addition was examined. Oxygen was depleted from the reaction mixture by extensive argon gassing. Exclusion of oxygen resulted in decreased TBARS production in plasma but not in standard solutions. High BHT concentrations resulted in a similar effect. At concentrations higher than 3 mmol/l BHT exclusion of oxygen had no additional effect. By measuring n-butanol extracts in a multititer plate reader this modified method was made suitable as a preliminary screening assay of human body fluids for lipid peroxidation.