High-pressure liquid chromatography (HPLC) grade methanol (Fisher, Hampton, NH), acetonitrile (The Lab Depot, Dawsonville, GA), and dichloromethane (DCM) (The Lab Depot, Dawsonville, GA) were used for all experiments. Deionized (DI) water was prepared with an on-site water purification system (Aqua Solutions Inc., Jasper, GA). Ammonium formate and formic acid (99%) were acquired from Sigma Aldrich (Pittsburg, PA). Isotopically labeled (²H₅) standards of cyclopropylfentanyl, 2-furanylfentanyl, acrylfentanyl, isobutyrylfentanyl, ocfentanil, and methoxyacetylfentanyl were purchased from Cayman Chemical (Ann Arbor, MI). Fentanyl, norfentanyl, and corresponding ²H₅ labeled standards as well as acetylfentanyl and corresponding ¹³C₆ labeled standard were purchased from Cerilliant (Round Rock, TX). Naloxone, naltrexone, heroin, 6-MAM, morphine, morphine-6-G, cocaine, and norcocaine were also purchased from Cerilliant. Norlofentanil and corresponding ²H₃ labeled standard were purchased from Toronto Research Chemicals (Toronto, Canada). Carfentanil, norcarfentanil, sufentanil, norsufentanil, corresponding ²H₅ labeled standards, and ¹³C₆-alfentanil were custom synthesized by Battelle (Columbus, OH). Pooled urine and pooled plasma along with individual urine and plasma reference samples were purchased from Tennessee Blood Services (Memphis, TN). This study does not meet the definition of human subjects as specified in 45 CFR 46.102 (f) as all urine and plasma samples were acquired from commercial sources with appropriate institutional review board approvals.

Three novel opioids and benzodiazapines (NOB) survey samples from the College of American Pathologists (CAP) were used to challenge our method. The samples, prepared by CAP in processed ovine blood, were evaluated in the same manner as human plasma samples used for method development.

CDC has contracted Cayman Chemical (Ann Arbor, MI) to manufacture and distribute the FAS Kit containing 200 micrograms each of 120 fentanyl analogs and metabolites analytical reference materials (Fentanyl Analog Screening Kit, FAS Kit). In addition, an expansion pack (Emergent Panel Version 1, FAS V1) was developed to contain 200 micrograms of an additional 30 synthetic opioid related compounds. Each compound in the FAS Kit and FAS V1 was provided in separate, individual vials. A list of all synthetic opioids and related compounds in the FAS Kit and FAS V1 can be found on the vendor’s website (https://www.caymanchem.com/forensics/faskit/). FAS kit development is explored in greater depth by Mojica et al. (Mojica et al, 2019)

Individual stock solutions of all analytes, purchased individually or provided in the FAS Kit and FAS V1, were prepared at 10 μg/mL in a mixture of methanol and DI water at a ratio of 3:2, respectively, with individual working solutions generated by diluting to 100 ng/mL in DI water. A 25 ng/mL internal standard (IS) working solution was created by a mixture of the isotopically labeled standards of fentanyl, carfentanil, acetylfentanyl, 2-furanylfentanyl, cyclopropylfentanyl, acrylfentanyl, sufentanil, ocfentanil, and methoxyacetylfentanyl. To form a 1 μg/mL quality control (QC) stock solution, 4-ANPP, acetylfentanyl, carfentanil, cyclopropylfentanyl, fentanyl, fluoroisobutyrylfentanyl, furanylfentanyl, methoxyacetylfentanyl, naloxone, and norfentanyl were diluted from the 10 μg/mL stock solutions. The QC stock solution was further diluted in pooled urine and plasma to create a positive QC Low (QCL) at 2 ng/mL and a positive QC High (QCH) at 15 ng/mL. In addition, an aliquot of pooled urine and plasma with no fortification was designated as a negative QC Blank (QCB).

A 200 μL aliquot of urine or plasma sample was pipetted into a 2 mL conical bottomed 96-deep well plate. IS working solution (25 μL) was added to the 96-deep well plate, followed by 175 μL of 0.1% v/v formic acid in DI water. The 96-deep well plate was sealed with adhesive foil and mixed at 1,000 rpm for 5 minutes (Eppendorf MixMate, Hauppauge, NY). The extraction was automated using a Biotage Extrahera (Charlotte, NC). The diluted sample was pipetted onto a Biotage ISOLUTE SLE+ 400 μL plate and given a 5 second burst of positive pressure air. The sample was allowed to absorb onto the media for 5 minutes, after which time 900 μL of DCM was applied to each sample well in the SLE plate. The DCM eluted without added pressure for 5 minutes into an empty 96-deep well plate. After 5 minutes, 0.7 bar of positive pressure was applied to the SLE plate to ensure the entire aliquot of DCM had eluted before a second 900 μL aliquot of DCM was added again to each sample well in the SLE plate. After 5 minutes had elapsed after the addition of the second 900 μL aliquot, a final 5 second burst of positive pressure was applied to the SLE plate. The collection plate was removed from the Extrahera and dried down with N₂ using a Porvair TurboVap (Ashland, VA) at a maximum temperature of 55 °C until dryness. The dried samples were then reconstituted in 100 μL of 78:22 10 mM ammonium formate in water : 0.1% v/v formic acid in acetonitrile. The 96-deep well plate was sealed with adhesive foil and shaken at 1,000 rpm for 5 minutes. The samples were then transferred into a 96-well PCR plate, heat sealed, and loaded into the instrument for analysis.

An Agilent Technologies 1290 Infinity II Liquid Chromatography (LC) system (Santa Clara, CA) with a 100 × 3.0 mm Phenomenex (Torrance, CA) biphenyl column kept at 50 °C was used for chromatographic separation. The column has a particle size of 2.6 μm and a pore size of 100 Å. For separation the eluents (A) 10 mM ammonium formate in DI water and (B) acetonitrile containing 0.1% v/v formic acid were used in the following gradient with a 700 μL/min flow rate: 78% A held for 0.5 min then reduced to 75% A over next 8.5 min. After 9 min A reduced to 70% over 2 min, then dropped to 60% at 11.01 min. From 11.01 min A was reduced to 55% over 1.99 min, and at 13 min A reduced to 5% over 0.60 min, at which it was held until chromatography completion at 16 min. The sample (15 μL) was injected, and the needle multi-washed with methanol containing 1% v/v formic acid and 82:18 10 mM ammonium formate in DI water : 0.1% v/v formic acid in acetonitrile before each injection.

Mass analysis was performed with an Agilent 6545 Q-TOF mass spectrometer in Auto MS/MS mode controlled using Agilent’s MassHunter Data Acquisition Version B.09.00. Analytes were ionized in positive mode electrospray ionization (ESI) using an Agilent Jet Stream source. The first 0.5 minutes and last 2 minutes of the chromatographic separation were diverted to waste. A capillary voltage of 3500 V and nozzle voltage of 1000 V was used for ESI, along with nebulizer and sheath gas (ultra-high purity nitrogen) at 350 °C to assist in ionization. For broadband MS analysis a mass range of m/z 100–1000 was analyzed a rate of 5 spectra/s and a time of 200 ms/spectrum. For each MS cycle, two precursors within m/z 200–600 and with at least 1000 counts abundance were automatically selected for MS/MS. Those precursors were then dynamically excluded for 0.1 minutes. To conserve cycle time, all IS compounds were placed on a static exclusion list, except fentanyl-D₅ which was used as a control to ensure sample viability. To ensure identification of library components, a list of the compounds in the library was used for preferential precursor selection. Precursors were isolated with a medium isolation width (~4 Da wide) and fragmented by collision-induced dissociation (CID) at 20 eV and 40 eV. The fragments were acquired at a rate of 3 spectra/s and a time of 333.3 ms/spectrum across m/z 50–1000. The instrument was externally calibrated daily with Agilent low concentration ESI tuning mix, and each analysis internally calibrated with purine and HP-0921 (hexakis(1H, 1H, 3H-tetrafluoropropoxy)phosphazine) from Agilent’s ESI-TOF reference mass solution kit.

The molecular formula of the synthetic opioids, fentanyl analogs, or other associated compounds found in the FAS Kit, FAS V1, or available in-house were entered into a personal compound database and library (PCDL) along with their calculated monoisotopic mass using MassHunter PCDL Manager B.08.00 (Agilent). To acquire mass fragmentation spectra, each individual compound stock solution was diluted to 25 ng/mL in DI water and analyzed in triplicate. The retention time (RT) was averaged and added to the PCDL alongside the fragmentation spectra. A table of all compounds in the in-house PCDL can be found in supplemental information (Table S1).

Chromatographic peaks were extracted in MassHunter Qualitative Analysis Workflows (Agilent, MassHunter Qualitative Analysis software version B.10) and identified by accurate mass and isotopic spacing using the Find-by-Formula algorithm against the in-house PCDL with a mass tolerance of ±5 ppm and RT tolerance of ±0.50 minutes. Only chromatographic peaks with a height greater than 7000 counts were extracted. Find by Formula’s match score is weighted based on mass (32.3%), isotope abundance (19.4%), isotopic spacing (16.1%), and RT accuracy (32.2%). The MS/MS spectra of the precursors identified by the Find by Formula were then used for library matching against the in-house PCDL, containing the reference CID spectra at 20 and 40 eV, using the “Identify Compounds” tool in Qualitative Analysis Workflows with an allowable mass error of ±10 ppm.

This method was validated for both urine and plasma following SWGTOX guidelines, as described in the following paragraphs (Scientific Working Group for Forensic, 2013).

Twenty replicate analytical runs of QCL, QCH, and QCB were analyzed to evaluate internal standard abundance, retention time, and library scores. These analytical runs were extracted and analyzed by two analysts, with no more than two replicates per day, over the course of 10 separate days.

Extraction efficiency and matrix effects were investigated for the QC compounds (Table 1). Data is only shown for 15 ng/mL but matched calculations for 2.0 ng/mL. Plasma extraction efficiencies ranged from 48.9 – 91.6%; urine extraction efficiencies ranged from 43.3 – 92.1%. The majority of the compounds had extraction efficiencies greater than 80% or above in both matrices. Only 4-ANPP and naloxone had lower extraction efficiencies, which could be attributed to their differences in chemical structures relative to fentanyl analogs. Matrix effects were below the SWGTOX guidelines of 25% with the exception of three compounds for urine and one for plasma (Table 1)(Scientific Working Group for Forensic, 2013). With the exception of fluoroisobutyrylfentanyl, the compounds with high matrix effects eluted at the extremes of the chromatographic run, which coincides with the elution of matrix components.

The LRL₁₀₀ was determined for the QC compounds (Table 1). LRL₁₀₀ for these compounds ranged from 0.25 – 1.00 ng/mL with the exception of norfentanyl and naloxone which were higher. In addition, LRL₁₀₀ were higher in urine than plasma. Compounds with lower extraction efficiencies, higher matrix effects, or a combination had higher LRL₁₀₀. The selected approach to determine LRL resulted in a conservative estimate to best describe method performance across time and variable conditions. Reported concentrations of fentanyl and fentanyl analogs following exposure have varied greatly, with fentanyl, carfentanil, acetylfentanyl, and furanylfentanyl concentrations ranging from 0.0102 ng/mL to 827 ng/mL in human matrices (Butler et al., 2018; Henderson, 1991; Martucci et al., 2018; Mochizuki et al., 2018; Shanks and Behonick, 2017; Sofalvi et al., 2017; Swanson et al., 2017). Although sensitivity may preclude the detection of all exposures due to delayed sample collection or opioid toxicity, this method has the capability to identify 174 fentanyl related compounds. This method can confirm the presence of analogs based on library match criteria, but the absence of an analyte cannot be definitively reported without characterization of the individual compound LRL.

Matrix interferences did not result in the positive identification of library compounds in the analysis of 50 individual urine samples and 50 individual plasma samples. However, fentanyl, acetylfentanyl, fluroisobutrylfentanyl, morphine, and norfentanyl were all positively identified in one of the individual plasma samples. Identification of fentanyl and acetylfentanyl was confirmed with a second method (data not shown). The addition of isotopically labeled standards or the metabolites of drugs commonly associated with fentanyl use did not result in any false positives.

Carryover was not observed in matrix blanks following a highly concentrated sample (100 ng/mL). Stability of processed samples was assessed over a period of 24 hours with no decrease in peak area counts.

Three quality control materials (QCL, QCH, QCB) were extracted and analyzed in duplicates over the course of 10 days. All QC analytes were positively identified in QCL and QCH across all runs, with the exception of norfentanyl, which had an LRL₁₀₀ above QCL concentration. No false positives were detected in the QCB.

In analysis of the three CAP evaluation samples, eight compounds were positively identified (Table 2) with library scores greater than 70 and mass error less than 5 ppm. A contemporaneous reference standard was analyzed within 24 hours for each positive identification and retention times were compared (Table 2) to confirm identification.

During the analysis of NOB-03, there were three possible isomeric identifications for a single peak: para-fluorofentanyl, meta-fluorofentanyl, and ortho-fluorofentanyl. As demonstrated from the chromatogram of all three standards and the CAP sample peak (Figure 4) the retention time of the sample most closely matches that of para-fluorofentanyl. With the similar retention times and library match we confirmed the peak as para-fluorofentanyl. Without the addition of the FAS Kit and the FAS V1 materials into this work’s newly implemented spectral database, the authors would not have known that the method would chromatographically separate the fluorofentanyl isomers ortho, meta, and para. Additionally, if the specimen had contained the ortho or meta isomers, and the laboratory was limited to a reference standard for only the para-fluorofentanyl, the method would not have been able to distinguish the specific fluorofentanyl isomer present, a potentially important piece of information critical to exposure surveillance. The application of the FAS Kit in this work, therefore, demonstrates the ability of the new analytical reference materials to expand opioid testing capabilities. The CAP NOB survey report, received after analysis, confirmed that para-fluorofentanyl was spiked into the sample, in agreement with our identification. Even with retention time tolerances and tight mass error criteria, this demonstrates the importance of analyzing known standards to positively identify unknown peaks, preferably within 24 hours as requested by a number of data reporting programs. In addition to the para-fluorofentanyl, all synthetic opioid-related compounds identified in table 2 were confirmed by the CAP NOB survey report to be found in the samples indicating there were no false positives. In addition, there were no false negatives, confirmed by the CAP NOB survey report.