Detection of illicit drugs in oral fluid from drivers as biomarker for drugs in blood

The aim of the study was to assess whether analysis of oral fluid can be used to identify individual drivers with drug concentrations in blood above 25 ng/mL for amphetamine and methamphetamine, 10 ng/mL for cocaine and 1.0 ng/mL for tetrahydrocannabinol (THC), which are the cut-off concentrations used in the European DRUID Project, by calculating the diagnostic accuracies when using the analytical cut-off concentrations in oral fluid as well as for the optimal cut-off concentrations. Paired samples of whole blood and oral fluid collected with the Statsure SalivaSampler were obtained from 4080 drivers in four European countries and analysed for amphetamine, methamphetamine, cocaine and THC using GC-MS or LC-MS. The vast majority (89%) were random drivers not suspected of drug-impaired driving. Receiver-Operating Characteristic analysis was used to evaluate the analytical results. The prevalence of drug findings above the cut-off concentrations in blood was 1.3% for amphetamine, 1.0% for methamphetamine, 0.6% for cocaine and 1.3% for THC. The cut-off concentrations in oral fluid that gave the highest diagnostic accuracy were for amphetamine 130 ng/mL (accuracy 99.8%), methamphetamine 280 ng/mL (accuracy 99.9%), cocaine 570 ng/mL (accuracy 99.6%), and THC 38 ng/mL (accuracy 98.3%). The proportion of false positives were 0.2%, 0.1%, 0.1% and 0.9%; and the proportion of false negatives were 0.1%, 0.0%, 0.3% and 0.8%, respectively, when using those cut-offs. The positive predictive values were 87.9%, 92.9%, 84.6% and 35.7% for amphetamine, methamphetamine, cocaine and THC, respectively. Analysis of concentrations of illicit drugs in oral fluid could not be used to accurately identify drivers with drugs concentrations above the selected cut-offs in blood in a cohort of drivers with low prevalence of drugs.