Identification of drugs in physiological fluids following on-line liquid chromatographic purification and analysis.

This paper describes the application of an automated algorithm for drug identification following the liquid chromatographic on-line separation and purification of drugs in physiological fluids. Two hundred fifty-seven (257) drugs and 45 drug metabolites were assembled for testing. Two internal standards were used which monitored different columns in the multi-step analytical separation. Retention reproducibility over a nine month period was 4 p. cent or less for most drugs. The algorithm for matching unknown peaks against the stored library utilized relative retention, lambda maxima, normalized spectral overlays, wavelength ratios and zero-intercepts from the second derivative in a combined forward-reverse search. Most of the drugs which could not be unambiguously identified were sedatives which eluted in the first minutes of the chromatogram. Diasteromers were partially resolved. Repetitive analysis of 200-400 micrograms/l of amphetamine, diphenhydramine, imipramine and morphine were evaluated to establish an absorption cutoff of 12 mAbs, above which at least 95% of peaks were correctly identified. This corresponded to a signal-to-noise ratio of 12 for the scanning UV detector which was evaluated.