The aim of the study was to examine the variation in ion suppression in ultra high pressure liquid chromatography tandem mass spectrometry (UHPLC-MS-MS) methods when using different blood collection devices. Three different methods measuring 18 antidepressants and antipsychotics in total were studied. The blood collection devices were all designed to activate clot formation. They were made of glass with or without silicone coating or plastic containing silicate particles, thrombin or polystyrene particles coated with kaolin. The blood collection devises Venoject and Venosafe were supplied from Terumo, S-monovette from Sarstedt, Vacuette from Greiner Bio-One and three BD Vacutainer serum tubes from BD. These seven different blood collection devices were used to withdraw blood from five healthy drug free donors (n=35) in random order. The samples were centrifuged and serum from each sample was aliquoted into two vials (A and B). The first vial (A) was used for post-column infusion experiments to visualize ion suppression. The second vial (B) was spiked with the 18 analytes in concentrations corresponding to the middle of the reference intervals of the analytes. The serum samples were analysed by UHPLC-MS-MS using three different gradients (Group I, II and III). The analytes in group I was measured on an Agilent 6460 mass spectrometer and group II and III were measured on an Agilent 6410 mass spectrometer both utilizing positive electrospray ionization. The experiments demonstrated significant differences in the absolute peak areas when using different blood collection devices. Similarly large differences between the analytes were observed. There was no difference between the donors or the two mass spectrometers. Using Venosafe the observed peak area was up to 76% lower than when using Venoject. The S-monovette and Vacuette tubes resulted in up to 49% and 33% lower peak areas, respectively. The analytes subjected to major ion suppression, when using the three blood collection devices, were 9-OH-risperidone, risperidone, duloxetine, and perphenazine. None of the three BD Vacutainer serum tubes caused any difference in the peak areas compared to the Venoject tubes. The observed ion suppression was substantiated by the post column experiments displaying unstable signal in the region of compound elution. Looking at the relative responses (peak area of the analyte/peak area of a stable isotope labelled internal standard), the use of different blood collection devices did not have the same significant impact on the outcome, thus stable isotope labelled internal standards for each analyte corrected for the ion suppression. In conclusion different blood collection devices used to produce serum can result in significant ion suppression. It is strongly recommended to include different blood collection devices in the validation protocol of new methods and to use stable isotope labelled internal standards for each analyte.