Semi-quantitative determination of designer benzodiazepines in serum by adaption of an automated LC-MSn screening approach

Abstract

Objectifs In 2012 benzodiazepine-type compounds started to extend the group of New Psychoactive Substances (NPS) sold on the internet as alleged legal recreational drugs. Initially, benzodiazepines like phenazepam and etizolam - which are still prescribed in some countries - were sold on the internet as recreational drugs. Subsequently, the group of so-called designer benzodiazepines was enlarged by compounds that either are active metabolites (e.g. norfludiazepam) or precursors (e.g. diclazepam) of known benzodiazepines or combine structural properties of different classical benzodiazepines (e.g. flubromazolam). Patents and scientific literature describe the synthesis and detailed results of animal model studies for more than a hundred different benzodiazepines, so it can be assumed that this sub-group of NPS will extend quickly in the future. The aim of this project was to develop an easy-to-use screening-method to detect and identify prescription and designer benzodiazepines and additionally obtain semi-quantitative information in a single run. Methods 1 mL serum sample was extracted using an established liquid-liquid extraction (LLE) protocol. Chromatographic separation was performed using a 13-minute formic acid/acetonitrile LC gradient and a Dionex Acclaim RLSC C18 100 × 2 mm column. An amaZon speed ion trap MS (Toxtyper ® , Bruker Daltonik) operating in ESI positive mode was used to generate MS2 and MS3 spectra according to a scheduled precursor list (SPL) triggered acquisition process. Compounds were identified by an in-house generated spectral library, containing retention time, MS and MS 2 /MS 3 information of currently 58 benzodiazepines. The linear range of the method was evaluated by fortifying blank human serum with benzodiazepine standards from 5 to 500 ng/mL. Semi-quantitative results were obtained using a single-point calibration (c = 50 ng/mL). Results The current spectral library contains 19 designer benzodiazepines, prescription benzodiazepines and metabolites. The method can easily be extended once new designer benzodiazepines emerge on the drug market or according to specific needs of the user. The LOD was 5 ng/mL except for six compounds only detected at concentrations above 10 ng/mL. Nifoxipam, known to have a poor ESI response and being highly instable in serum, was the only compound that could not be detected at practically relevant concentrations in serum. For each analyte, a linear calibration range (calLow to calHigh) was determined and calculated concentrations within this range are reported as semi-quantitative result in the automatically generated report. Findings below or above the linear range are reported as ‘ calHigh’, respectively. Semi-quantitative results were found to vary between ± 20 and ± 40% at the lower end of the calibration range and ± 10 to ± 25% at medium and high concentrations. Conclusion The presented method allows automated identification and semi-quantitative determination of 58 benzodiazepines, including 18 designer benzodiazepines. LOD in serum allow the detection below the desired or at the concentration needed for the majority of medical benzodiazepines, making the screening applicable for clinical research and forensic analysis. Semi-quantitative analysis enables a quick toxicological research evaluation of the results and helps to decide on the analytical strategy in case work. Although this approach requires a more time consuming sample preparation when compared to routine immunoassays, unambiguous identification and semi-quantitative determination of compounds also offers more detailed information.