Ricerc@Sapienza

In the field of forensic toxicology, the use of non-destructive and easy-to-use analytical techniques deserves remarkable attention, especially in those situations involving public health and security. In addition, the miniaturization
and portability of one-touch devices for the detection of specific threats is required more and more.
In this study, a novel on-site MicroNIR/Chemometric platform was developed to perform a real-time prediction
of cocaine and its metabolites in non pre-treated oral fluid.
Simulated oral fluids were prepared in water in order to calibrate the instrumental response and the matrix
effect was consequently evaluated by processing spiked oral fluids collected from volunteers. The procedure was
optimized using a proper experimental design taking into account the equilibrium between cocaine and benzoylecgonine
in the range 10–100 ng-ml and validated by comparing results with the reference official method
(GC-MS).
The developed method was statistically able to discriminate oral fluid samples containing cocaine from 10 to
100 ng/ml and demonstrated to be not affected by the variability of the matrix as all the blank samples of
different volunteers (smokers and non smokers, assuming caffeine, sugars, chewing-gum or alcohol) as well as
spiked oral fluids were correctly predicted by the model. In addition, results from six real samples confirmed the
feasibility of the miniaturized platform to provide a correct identification of cocaine abuse and to propose the
MicroNIR as innovative personal screening system to prevent accidents and in cases involving workplace surveillance.