Automated and Fully Validated High-Throughput LC-MS/MS Assay for Analyzing Multiple Drugs of Abuse in Oral Fluids Using Novel Features in Sample Preparation and Chromatographic Conditions

Oral fluid sampling offers advantages over other biological matrices, mainly due to its noninvasive procedure avoiding privacy intrusion. The fully automated sample preparation procedure is based on salting-out assisted liquid–liquid extraction (SALLE) combined with high-efficiency LC-MS/MS methods for both screening and confirmation of 37 drugs and incorporates novel features enabling direct injection of acetonitrile extracts into an innovative chromatographic system. The methods' drug panel includes opioids, benzodiazepines, benzodiazepine-like drugs, cannabinoids, and stimulants. A full method validation was performed using OF/buffer from Greiner Bio-ONE International and Quantisal saliva collection devices. The validation included assessments of linearity, sensitivity, precision, accuracy, extraction recovery, matrix effects, process efficiency, stability, and carryover. All compounds demonstrated linearity across the concentration range 1–25 ng/mL, with R² ≥ 0.99. Both methods' limit of detection ranged between 0.001 and 0.03 ng/mL, and the limit of quantification ranged between 0.02 and 0.09 ng/mL. Precision was ≤ 14.8% for screening and ≤ 8.5% for the confirmation method. Accuracy was ± 13.6% for screening and ± 8.7% (except at 0.5 and 1 ng/mL, where it was ± 25.3% and ± 17.6%, respectively) for the confirmation method. Extraction recoveries ranged from 40.0% to 95.1%, except for hydromorphone (27.4%) and morphine (34.4%). Although matrix effects were observed for a large number of compounds to varying degrees, they were largely compensated for by the use of deuterium- and ¹³C-labeled internal standards (IS). IS-corrected overall process efficiency ranged from 100.7% to 119.1% with precision (CV%) ≤ 10.8% for all compounds. Spiked calibrators and QC samples in OF were stable in autosampler for up to 72 h and in the freezer for 3 days. Methanol working solutions were stable for 6 months. No significant carryover was observed. The methods have been successfully implemented in the routine analysis of approximately > 1000 samples per month since March 2024.