Evaluation of the reproducibility and repeatability of GCMS retention indices and mass spectra of novel psychoactive substances

Abstract

Novel psychoactive substances (NPSs) are synthetically derived compounds designed to mimic the effects of other illicit drugs. An endless cycle of NPSs continually reach the drug market due to limitations in drug legislation creating two problems – non-availability of standards and multiple isobars indistinguishable even to high resolution mass spectrometry. Lacking standards, laboratories rely on published gas chromatography-mass spectrometry (GCMS) spectra to postulate compound identification. The ability to make such identifications provides vital investigative data and may help uncover new structural variants. However, central to such identifications is an understanding of how much spectral patterns vary across laboratories; the assumption that variation is minimized through the use of a standard tuning compound (PFTBA) has yet to be rigorously tested and demonstrated. In the present study, the variation of mass spectral fragmentation patterns characterized by percent relative abundance was evaluated using three GCMS instruments and 16 NPS compounds. The variation of retention indices was also studied. Retention indices showed <0.35% variation across all systems; mass spectral data showed much greater variation across all compounds and systems. Tuning frequency was one factor correlated with small variation. A means comparison of ion relative abundance did not clearly identify general patterns. Changes in identification of base ion peak was seen with three compounds. Base peak shifting on a single instrument, i.e. one instrument produced multiple base peak ions over the course of the study, generated a ripple effect, where variation across all relative abundances in the effected spectra increased. However, base peak shifting between instrument, i.e. each instrument produced a single base peak but the ion identity was not necessarily the same, did not produce these ripple effects. Either shifting base peak scenario had the potential to skew compound identification during untargeted analyses, which are common in a forensic drug chemistry setting.