Atti del XXIX congresso della divisione di chimica analitica della Società Chimica Italiana

Novel psychoactive substances (NPS) are a group of substances, mainly of synthetic origin, characterised by toxicological properties extremely dangerous for the users, often unaware about the real threat related with these substances. The aim of this study was the creation of a robust and versatile suspect screening method for traditional drugs and NPS, starting from a target method encompassing 97 analytes. The target compounds belong to different classes including natural and synthetic cannabinoids, cathinones, alkaloids, phenethylamines, arylcyclohexylamines, opioids and tryptamine. For this purpose, an UHPLC system coupled to a hybrid triple quad-linear ion trap MS/MS mass spectrometer was used. Target MS detection was carried out by advanced scheduled MRM (AsMRM) algorithm; a suspect screening method was then created by adding new MRM transitions of other NPS, not included in the target list. The molecular ions and fragments for Q1 and Q3 settings were obtained from HighResNPS [1] and the retention time was predicted by using a QSRR specifically developed. The prediction model used was a MLR between molecular descriptor and rT, they are then normalised, correlated and the best used in the calculation of the prediction equation. IDA acquisition mode that coupled an AsMRM as survey scan and an EPI as dependent acquisition was used. With this strategy, screening capabilities were improved and putative identification of different drugs by library matching may be possible even if standards are not available. Calculated LODs ranged between 0.007 ng/mL for 25C-NBOMe and 0.4 ng/mL for morphine. Oral fluid was considered for testing the method in a biological matrix and excellent results in terms of accuracy, precision and matrix effect were obtained. As a proof of concept, the performance of the suspect method was evaluated by analysing a mixture containing 21 reference standards not included in the initial dataset; 19 out of 21 analytes were correctly detected. Rt were generally in accordance with experimental Rt, with an average shift of about 13%. Finally, the suspect screening method was applied to four seizures. MXPR and 3-MMC were identified in these samples, showing that this approach is suitable to detect NPS even in real samples.