A simple automated procedure for the detection and identification of peaks in gas chromatography--continuous scan mass spectrometry. Application to systematic toxicological analysis of drugs in whole human blood.

Gas chromatography-mass spectrometry (GC-MS), which combines the separation power of GC with the power of MS for the identification of unknown compounds, possesses high potential in systematic toxicological analysis (STA). Different factors, however, do not allow this potential to be fully exploited. Between them, the low selectivity of the mass spectrometer operating in continuous scan plays a critical role, in many cases precluding the possibility of selecting mass spectra in the total ion chromatogram (TIC) sufficiently clean for a positive identification by library search, even when using reverse search algorithms. Moreover, the large amount of information contained in GC-MS data file that results from the analysis of a biological extract makes the efforts of manual search almost useless and requires the availability of reliable methods for the automated detection and identification of peaks in a TIC. In this paper, a simple procedure that improves the performance of a bench-top GC-MS system in the purification of mass spectra of coeluting compounds and that can be easily combined with the automated processing of a GC-MS data file is described. It is based on the subtraction of the intensities of successive pairs of scans in the TIC, on the detection of positive and negative peaks in the transformed chromatograms, and on the search of the corresponding background-subtracted electron ionization mass spectra against reference libraries. In order to evaluate the proposed procedure, GC-MS data files obtained for the analysis of extracts of blank whole blood spiked with more than 100 drugs, poisons, and their metabolites at a concentration of 0.5 mg/L were used. Compared with the search of the raw TICs, the proposed procedure increased the number of identified substances and, in many cases, obtained higher match quality values for identification.