Comparison of human metabolome changes identified in a placebo-controlled amphetamine administration study versus those using forensic toxicology routine data

Metabolome studies in forensic toxicology focus on the search for endogenous biomarkers changed by, e.g., drugs of abuse. However, placebo-controlled studies, the ideal study design, in humans are scarce for ethical reasons. Thus, the idea of using routine samples became popular, although confounding factors cannot be controlled. To systematically evaluate the use of routine samples for metabolomics, a comparison between a placebo-controlled amphetamine study in humans (A, n pos =18, n neg =18) to routine samples either positive or negative for amphetamine, prepared and analyzed over six months (re-evaluated, B, n pos =28, n neg =35) and prepared and analyzed within a single analytical batch (re-extracted, C) was performed. Samples were analyzed using untargeted liquid chromatography-tandem-mass-spectrometry. Comparison was conducted on feature level and based on significance (p- and fold-change-values). Only 3 features were significant in A, B, and C, and 2 were identified as amphetamine-(fragments). All 31 significant features from A were present in B and C; however, only 11 (36%) and 4 (13%) of them were significant mainly because of higher variation. Still, other significant features were found in routine samples (B/C). In conclusion, routine samples are generally suitable for detecting differences in the metabolome, even if they do not match those of a controlled study.