Comparison of measured and recommended acceptance criteria for the analysis of seized drugs using Gas Chromatography-Mass Spectrometry (GC-MS)

Gas Chromatography-Mass Spectrometry (GC-MS) is a widely used analytical technique that has become a critical tool in many industries, including forensic science. Many governing bodies provide recommendations for the uncertainty of measurement for chemical substance identification, but existing guidelines often do not provide the numerical data to support the basis for their acceptance criteria. The guidelines therefore appear anecdotal and, if they are not continually updated, may not reflect modern instrument capabilities. This study provides data, with detailed interpretations, to assess the magnitudes and sources of measurement uncertainties of typical GC-MS data, as commonly practiced. Data analysis was conducted using 13 different drug standards from three different laboratories using five different GC-MS setups. The laboratories were not prescribed a set of instrumental parameters, but rather were asked to submit the parameters ordinarily practiced within their respective laboratories. An expanded uncertainty of two times the relative standard deviation (2 sigma) of replicate measurements was used to report the uncertainty of measurement for the retention time and relative ion abundance measurements made on each set-up. The retention time acceptance criteria currently recommended by many agencies are on the magnitude of +/- 2%, but such criteria are much wider than the measured within-week or within-month 2 sigma values, which are actually on the magnitude of +/- 0.20%. The measured uncertainties of relative ion abundances are similar to recommended acceptance criteria, but a careful assessment shows that ion abundances are not independently variable within a spectrum. Some ion abundances correlate with correlation coefficients (R squared) that exceed 0.9. Acceptance criteria for the GC retention time measurements should therefore be stricter than most of the current recommended guidelines. The application of tighter acceptance criteria would provide: 1) an evidenced-based, statistical reason for making drug identifications; and 2) fewer type I errors (false positives) in seized drug analyses than provided by existing standards. (C) 2018 Published by Elsevier B.V.