Résumé : The use of amphetamine and 'ecstasy' (MDMA) has increased exponentially in many European countries since the late nineties, leading to a rapid growth in the number of clinical and forensic analyses. Therefore, a rapid screening procedure for these substances in biological specimens has become an important part of routine toxicological analysis in forensic laboratories. The objective of this study was to evaluate the Cozart® amphetamine enzyme-linked immunosorbent assay (ELISA) for the screening of plasma samples and oral fluid samples (collected with the Intercept® device). Authentic plasma samples from drivers (n = 360) were screened, using an 1:5-fold dilution. True positive, true negative, false positive and false negative results were determined relative to the in-house routine GC-MS analysis. Samples consisted of 144 amphetamine-only positives, 141 MDMA/MDA-only positives, and 74 negatives when using the limit of quantitation as the cut-off level for confirmation (10 ng/mL). Using these results, receiver operating characteristic (ROC) curves were generated and optimal cut-off values for the screening assay were calculated. Analysis showed that the ELISA is able to predict the presence of either amphetamine or*MDMA/MDA (*MDMA as its metabolite MDA) in plasma samples with 98.3% sensitivity and 100% specificity at a cut-off value of 66.5 ng/mL d-amphetamine equivalents. A similar analysis was conducted on 216 oral fluid specimens collected from a controlled double blind study. Subjects received placebo or a high (100 mg) or low (75 mg) dose of MDMA. Oral fluid samples were collected at 1.5 and 5.5 h after administration. Combined results of the analysis of the high and low dose oral fluid samples indicated a screening cut-off of 51 ng/mL d-amphetamine equivalents with both a sensitivity and specificity of 98.6% (using a LC-MS/MS confirmation cut-off of 10 ng/mL). In conclusion, these data indicate that the Cozart® AMP EIA plates constitute a fast and accurate screening technique for the identification of amphetamine and MDMA/MDA positive plasma samples and oral fluid specimens (collected with Intercept®). It should be emphasized that method validation should be performed for each type of biological matrix. © 2005 Elsevier Ireland Ltd. All rights reserved.