Résumé : Phosphatidylethanol species (PEths) are promising biomarkers of alcohol consumption. Here, we report on the set-up, validation, and application of a novel UHPLC-ESI-MS/MS method for the quantification of PEth 16:0/18:1, PEth 18:1/18:1, and PEth 16:0/16:0 in whole blood (30 μL) and in venous (V, 30 μL) or capillary (C, 3 punches (3 mm)) dried blood spots (DBS). The methods were linear from 10 (LLOQ) to 2000 ng/mL for PEth 16:0/18:1, from 10 (LLOQ) to 1940 ng/mL for PEth 18:1/18:1, and from 19 (LLOQ) to 3872 ng/mL for PEth 16:0/16:0. Extraction efficiencies were higher than 55 % (RSD < 18 %) and matrix effects compensated for by IS were between 77 and 125 % (RSD < 10 %). Accuracy, repeatability, and intermediate precision fulfilled acceptance criteria (bias and RSD below 13 %). Validity of the procedure for determination of PEth 16:0/18:1 in blood was demonstrated by the successful participation in a proficiency test. The quantification of PEths in C-DBS was not significantly influenced by the hematocrit, punch localization, or spot volume. The stability of PEths in V-DBS stored at room temperature was demonstrated up to 6 months. The method was applied to authentic samples (whole blood, V-DBS, and C-DBS) from 50 inpatients in alcohol withdrawal and 50 control volunteers. Applying a cut-off value to detect inpatients at 221 ng/mL for PEth 16:0/18:1 provided no false positive results and a good sensitivity (86 %). Comparison of quantitative results (Bland-Altman plot, Passing-Bablok regression, and Wilcoxon signed rank test) revealed that V-DBS and C-DBS were valid alternatives to venous blood for the detection of alcohol consumption.