Drug overdose mortality has risen dramatically in North America over the last two decades. The overdose crisis is now considered to be in its “Fourth Wave,” characterized by high mortality driven by synthetic opioids replacing traditional illicit drugs and augmented by increasing stimulant-involved deaths. In 2020 and 2021 in San Francisco County, cumulative deaths from overdose were almost double those from COVID-19, and an increase in fentanyl-involvement in deaths among individuals without prior evidence of opioid use in San Francisco suggests increased unintentional fentanyl exposure in the city.
Drug checking, the use of chemistry techniques to identify components present in a drug sample, has gained traction as one harm reduction intervention to mitigate overdose. Drug checking in the US has so far been limited to FTIR spectroscopy and fentanyl immunoassay test strip technologies to provide onsite service. San Francisco launched its own FTIR and test strip service in June 2022, with the service augmented with offsite research laboratory confirmation of all drug samples analysis performed by untargeted Liquid Chromatography-Quadrupole Time of Flight-Mass Spectrometry (LC-QTOF-MS). This work demonstrates the advantages of this onsite-offsite hybrid model of drug checking, which provides all the benefits of a point-of-service test for people who use drugs (PWUD), while gaining superior chemical analysis to better comprehend the composition of the illicit substance supply. Improved knowledge of the local illicit substance supply can be utilized to inform public health planning and toxicological case response.
To demonstrate the effectiveness of utilizing a clinically validated LC-QTOF-MS comprehensive drug screen method to analyze street drug products to evaluate, confirm, and augment point of service drug checking efforts.
All drug samples submitted for point-of-service drug checking were delivered to the clinical lab dissolved in 1 mL MeOH at approximately 1 mg/mL to circumvent transportation of consumable drug material. All samples were diluted to approximately 10 mcg/mL in mobile phase and analyzed using a method previously validated for clinical testing. Briefly, chromatographic separation was performed using a C-18 column with a 10-minute gradient from 2%-100% organic. Data was collected on a SCIEX TripleTOF®5600 operating in positive-ion mode using a TOF-MS survey scan with IDA-triggered collection of high-resolution product ion spectra (20 dependent scans). Data was analyzed using an in-house library containing >5000 small molecules including >150 fentanyl analogs. Calibrator mixes were made for 25 compounds of interest to enable “semi-quantitation” in the drug products. Accurate quantitation is limited by weighing errors at the point-of-service and issues with drug product dissolution. Results were available to point-of-service drug checking staff within three days of sample intake for the purposes of informing delivery of service and to communicate confirmation results to returning service participants.
From June to November 2022, 188 drug products were evaluated at the point-of-service using FTIR spectroscopy and immunoassay lateral flow test strips for fentanyl and benzodiazepines followed by analysis using the LC-QTOF-MS method. The LC-QTOF-MS detected an additional 36 compounds in the samples that were missed by point-of-service technologies. Using the LC-QTOF-MS method as the “gold standard,” the sensitivity and specificity of the FTIR was 36.18% and 98.84% respectively for all analytes. Sensitivity and specificity of the FTIR for fentanyl only was 58.70% and 100.00% respectively, improving to 93.48% and 98.59% respectively when fentanyl test strip results were considered. Notably, the FTIR missed all benzodiazepine compounds detected by LC-QTOF-MS (n=4).
Confirmation by LC-QTOF-MS was originally planned for every sample only during the pilot period of the point-of-service program, but due to the overwhelming advantages provided by confirmation, confirmation for all samples has been extended indefinitely.
As determined by LC-QTOF-MS, samples expected to be methamphetamine (n=72) rarely contained other drug classes (n=2, 4.17%). Samples expected to be cocaine (n=42) contained unexpected drug classes more often, (n=10, 23.81%), but 9 of these 10 samples contained only lidocaine or levamisole, established cocaine adulterants. Samples expected to be fentanyl (n=37) were frequently identified as complex mixtures of compounds from a variety of drug classes (n=24, 64.86%), with semi-quantitation identifying widely varied and unpredictable potency of fentanyl in the samples ranging from 0.0092% to 46.37%. Samples expected to be heroin (n=10) also showed frequent contamination with fentanyl (n=3, 30.00%), conflicting with established thought that fentanyl is unlikely to be mixed with West Coast tar heroin. Detection of xylazine, deschloro-etizolam, flubromazolam, bromazolam, and butyryl fentanyl by confirmation but not by point-of-service demonstrate the added benefit of utilizing mass spectrometry to improve public health surveillance efforts to identify drug supply changes to inform overdose crisis response efforts.
Confirmation of point-of-service drug checking by LC-QTOF-MS provides deeper comprehension of the illicit substance supply in a geographic region, aiding both harm reduction and traditional clinical practice. This information facilitates improved confidence in the capability and limitations of point-of-service drug checking, improves drug checking as public health surveillance intervention, and provides crucial context to interpret comprehensive drug screen results for clinical cases.