Separation and Quantification of Superwarfarin Rodenticide Diastereomers—Bromadiolone, Difenacoum, Flocoumafen, Brodifacoum, and Difethialone—in Human Plasma

Abstract

Background

Superwarfarins, second-generation long-acting anticoagulant rodenticides, are 4-hydroxycoumarin analogues of warfarin that contain a large hydrophobic side chain. These compounds contain two chiral centers and are synthesized for commercial use as two pairs of diastereomer.

Objective

To support studies of superwarfarin pharmacokinetics and other efforts to improve clinical care for poisoning victims, a quantitative assay was developed for the measurement of diastereomer of bromadiolone, difenacoum, flocoumafen, brodifacoum, and difethialone in human plasma.

Method

Based on ultrahigh-pressure liquid chromatography-tandem mass-spectrometry (UHPLC-MS/MS), this method was validated according to U.S. Food and Drug Administration (FDA) guidelines. Sample preparation involved simple protein precipitation followed by reversed phase UHPLC, which resolved all five pairs of cis/trans diastereomer in less than 10 min. Superwarfarins were measured using negative ion electrospray followed by selected-reaction monitoring on a triple quadrupole mass spectrometer.

Results

Calibration curves covered 3–4 orders of magnitude with linear regression coefficients of >0.999. The lower limits of quantitation were from 0.013 to 2.41 ng/mL, and intra-day and inter-day accuracy and precision coefficients of variation were <12%.

Conclusions

A 10-min UHPLC-MS/MS assay was developed and validated for the separation and quantitative analysis of the pairs of diastereomer of five superwarfarins in human plasma.

Highlights

This method was used to identify and measure superwarfarins and their cis/trans diastereomers in plasma obtained from patients treated for coagulopathy following consumption of contaminated synthetic cannabinoid products.

Based on reports of coagulopathy in cattle eating moldy yellow sweet clover (Melilotus officinalis), the 4-hydroxy coumarin anticoagulant dicoumarol was isolated in 1939 (1). Recognizing the therapeutic capability of an orally active anticoagulant, dicoumarol was used clinically to prevent thrombosis (2) until 1954, when dicoumarol was replaced by a more potent synthetic derivative, warfarin (3). Warfarin was also used as a first-generation anticoagulant rodenticide and was considered effective until the 1960s, when rodents developed resistance (4). To overcome resistance, non-polar analogues of warfarin, known as long-acting anticoagulant rodenticides or superwarfarins, were introduced in the 1970s and 1980s (5, 6), and remain in use today.

Superwarfarins, bromadiolone (BDL), difenacoum (DFC), flocoumafen (FCF), brodifacoum (BDF), and difethialone (DFT) retain the 4-hydroxy coumarin structure of dicoumarol and warfarin but contain an extra lipophilic side-chain (Figure 1). This substituent increases anticoagulation potency 100-fold compared to warfarin, and prolongs half-life to 20–30 days (7, 8). Depending upon the initial dosage, toxic levels of superwarfarins can persist for up to 1 year (9).

Figure 1.

Chemical structures of bromadiolone (BDL), difenacoum (DFC), brodifacoum (BDF), flocoumafen (FCF), difethialone (DFT), dicoumarol, and warfarin. The 4-hydroxy coumarin scaffold, highlighted in bold, is conserved among the superwarfarins, dicoumarol and warfarin. Note that each superwarfarin contains two asymmetrical carbons forming four isomers (two pairs of diastereomers).

Like dicoumarol and warfarin, superwarfarins function by competitive inhibition of vitamin K 2,3-epoxide reductase (VKORC1), which disrupts the vitamin K cycle and prevents activation of clotting factors (10–12). Although intended as rodenticides, approximately 10,000 humans, mainly children, are poisoned each year by superwarfarins in the United States (13–18). Therapeutic interventions following superwarfarin poisoning include intravenous vitamin K₁ (VK₁), fresh frozen plasma, red cell transfusion, or a prothrombin protein complex to restore coagulation followed by maintenance doses of oral VK₁ (7).

Warfarin contains one chiral carbon and is often synthesized as a racemic mixture. Superwarfarins contain two asymmetric carbons and are synthesized as racemates, thereby forming two pairs of diastereomers (Figure 1) (19, 20). The S-enantiomer of warfarin inhibits VKORC1 three to five times more effectively than does the R-enantiomer, a R- and S-warfarin are metabolized differently (21, 22). In order to probe possible pharmacokinetic differences between superwarfarin diastereomers, suitable analytical methods are required. Separation of diastereomers can facilitate the development of safer rodenticides, for example, when one pair of superwarfarin diastereomers is selectively less toxic toward non-target species (8, 23). The separation of superwarfarin diastereomers can also benefit forensic investigations by providing information regarding the source of a particular rodenticide (24, 25); this is based on the observation that the ratio of cis: trans diastereomers varies depending on the synthetic reagents and route used during synthesis (19, 26–28).

Approaches to separate and measure superwarfarin rodenticides in animal and human plasma, serum and tissues have utilized thin-layer chromatography (29, 30), gas chromatography-mass spectrometry (31, 32), liquid chromatography with UV or fluorescence detection (23, 29, 33–48), liquid chromatography-mass spectrometry (LC-MS) (49), and LC-tandem mass spectrometry (LC-MS/MS) (50–63). Among these previous methods, eight included all five superwarfarin rodenticides (34, 50, 52, 55–57, 62, 63), but only three of those are validated for human blood (55, 62, 63), and none of these three resolve superwarfarin diastereomers or have the calibration range to measure concentrations above 200 ng/mL.

Here we report the development and validation of a rapid and sensitive method based on ultrahigh-pressure liquid chromatography (UHPLC)-MS/MS for the separation and quantitative analysis of diastereomers of BDL, DFC, FCF, BDF, and DFT in human plasma. This method was then applied to the measurement of superwarfarins in human plasma from patients who had ingested or smoked contaminated synthetic cannabinoid products in 2018 (64–66).

Experimental

Chemicals and Reagents

LC-MS grade acetonitrile and formic acid as well as superwarfarins BDL, DFC, FCF, and BDF were purchased from Sigma-Aldrich (St. Louis, MO). The superwarfarin DFT and internal standard BDF-d₄ were purchased from Toronto Research (Ontario, Canada) (Figure 1). Blank human plasma was purchased from Lampire Biological Laboratories (Pipersville, PA). Whole blood samples from subjects were obtained from the OSF Saint Frances Medical Center (Peoria, IL). Purified water was prepared using a Milli-Q water purification system (Millipore, MA).

Plasma Preparation

Whole blood samples, which had been collected in vacutainers and stored at –80°C until processing, were thawed, vortexed and centrifuged at 20,000 × g for 20 min at 4°C. Protein precipitation was performed by adding 400 µL of 10% methanol in acetonitrile containing BDF-d₄ (37.5 ng/mL) into 100 µL of plasma. The samples were vortexed for 5 min, centrifuged at 20 000 × g for 15 min at 4°C and 400 µL of supernatant were dried under vacuum. The residue was reconstituted with 50 µL of 50% (v/v) aqueous acetonitrile, vortexed for 5 min and centrifugated at 20 000 × g for 15 min at 4°C immediately before analysis using UHPLC-MS/MS.

Preparation of Calibration Standards and QC Samples

Stock solutions of BDL, DFC, FCF, BDF, DFT, and BDF-d₄ were prepared in DMSO (1 mg/mL) and stored at –20°C. Working solutions at concentrations ranging from 0.625 ng/mL to 40.96 µg/mL in 50% (v/v) aqueous acetonitrile for all five analytes were prepared from DMSO stock solutions using serial dilution and stored at –20°C. Calibration standards were prepared by spiking 95 µL of blank plasma with 5 µL of working solution followed by extraction as stated above. QC samples were prepared identically to the standards at the lower limit of quantitation (LLOQ), low, medium and high concentrations. The LLOQ was defined as a signal-to-noise ratio (S/N) ≥10, the low concentration QC was three-fold greater than the LLOQ, the medium concentration QC was 15% of the upper limit of quantitation (ULOQ), and the high concentration QC was set to 80% of the ULOQ for each analyte. The lower limit of detection (LLOD) was defined as a S/N of three.

Chromatography

Separations were carried out using a Shimadzu (Kyoto, Japan) Nexera UHPLC system equipped with a Shimadzu Nexcol C₁₈ (50 × 2.1 mm, 1.8 µm) column. The solvent system consisted of water (A) and acetonitrile (B) both containing 0.01% formic acid. A linear gradient at a flow rate of 0.6 mL/min was used as follows: 0.0 min, 30% B; 0.5 min, 35% B; 0.51 min, 61% B; 8.3 min, 68% B; 8.5 min, 68% B; 8.51 min, 30% B; 10.0 min, 30% B. The column temperature was 50°C and the injection volume was 10 μL.

Tandem Mass Spectrometry

Superwarfarins were measured using a Shimadzu LCMS-8060 triple quadrupole mass spectrometer operated at unit resolution and with negative ion electrospray. Nitrogen was used as a nebulization gas at a flow rate 3.0 L/min, heating gas at 10 L/min and a drying gas at 10 L/min. The interface temperature was 300°C, the desolvation line was 250°C and the heat block temperature was 400°C. Superwarfarins were measured using selected reaction monitoring (SRM) following collision-induced dissociation with argon gas at 230 kPa. SRM transitions for MS/MS quantification (quantifier) and for QC (qualifier) for each analyte are shown in Table 1.

Table 1.

UHPLC retention times, MS/MS collision energies (CE) and precursor/product ion selected reaction monitoring (SRM) transitions for superwarfarin diastereomers

Superwarfarin	Retention time, min	Precursor ion	Quantifier product ion		Qualifier product ion
		m/z	m/z	CE, V	m/z	CE, V
cis-BDL	2.2	527	250	37	93	41
trans-BDL	2.3	—	—	—	—	—
cis-DFC	3.1	443	293	34	93	50
trans-DFC	3.5	—	—	—	—	—
cis-FCF	4.2	541	382	25	161	37
trans-FCF	4.8	—	—	—	—	—
cis-BDF	4.9	523	135	37	81	40
trans-BDF	5.4	—	—	—	—	—
cis-DFT	8.0	537	79	45	151	38
trans-DFT	8.4	—	—	—	—	—
cis-BDF-d4	4.8	527	139	38	97	66
trans-BDF-d4	5.3	—	—	—	—	—

Calibration Curve Linearity

Calibration curves were constructed using 13 calibration standards in duplicate for each analyte. The natural log of the peak area ratio of analyte to internal standard was plotted against the natural log of the analyte concentration. The linearity of each calibration curve was determined using least squares linear regression analysis, without a weighing factor. Data analysis was carried out using Shimadzu Lab Solutions software (Ver. 5.95).

Selectivity and Matrix Effect

Defined as the extent to which components of the matrix might interfere with the SRM signal of each superwarfarin, the selectivity of the method was evaluated by analyzing six different lots of blank human plasma. Any peak detected at the retention time of a superwarfarin with an area greater than 20% of the analyte at the LLOQ or 5% of the internal standard was considered significant interference. Possible matrix effects, defined as the combined influence of all the non-analyte components of the analyte signal, were assessed by comparing the peak areas of spiked blank matrix and spiked neat injection solvent at low and high analyte concentrations. Deviation in peak area >15% compared with the neat injection solvent would be considered significant.

Precision and Accuracy

Three separate analyses on three consecutive days were performed to evaluate repeatability (intra-day variability), intermediate precision (inter-day variability) and accuracy. Each analysis consisted of a calibration curve and five QC samples at the LLOQ, low, medium, and high concentrations. Repeatability and intermediate precision were defined as the relative standard deviation (RSD) of replicate QCs at each concentration. Accuracy was defined as the percent difference between the mean of replicate QCs compared with the true nominal value at each concentration.

Recovery

The recovery of each superwarfarin from human plasma was determined at low, medium and high concentrations after protein precipitation by comparing the peak area ratios (analytical peak area:internal standard peak area) of experimental to theoretical samples. Experimental samples consisted of standards spiked into blank plasma, processed and reconstituted with neat injection solvent containing internal standard. Theoretical samples consisted of neat solvent spiked into blank plasma, processed and reconstituted with standards in the injection solvent containing internal standard.

Stability

Short-term stabilities of BDL, DFC, FCF, BDF, and DFT at low and high concentrations were assessed at room temperature on the benchtop and at 4°C in the autosampler over 24 h. Long-term stabilities at low and high concentrations were assessed for samples stored at –20 and –80°C for 30 days.

Results and Discussion

LC-MS/MS Method Development

To achieve separation of cis- and trans- diastereomers of the five superwarfarins, reverse-phase HPLC and UHPLC columns were evaluated along with various mobiles phase compositions. Using a superficially porous Agilent Poroshell EC-C₈ column (50 × 2.1 mm, id 2.7 µm), diastereomers of DFC, FCF, BDF, and DFT could be resolved in a single run but not those of BDL. However, a fully porous Shimadzu Nexcol C₁₈ column (50 × 2.1 mm, id 1.8 µm) was found to facilitate the separation of the cis/trans diastereomers of all five superwarfarins.

Using a gradient of increasing acetonitrile in water, different concentrations of formic acid were compared, and although 0.1% formic acid provided the best peak shape, the negative ion electrospray tandem mass spectrometer response was slightly suppressed. Therefore, 0.01% formic acid was used in the mobile phase to provide the optimum balance of peak shape, selectivity and MS/MS sensitivity. Using this mobile phase composition, a shallow 10-min linear gradient was found that provided not only chromatographic resolution of BDL, DFT, BDF, DFC, and FCF (with retention times from 2.2 to 8.4 min) (Table 1) but also separation of each cis/trans pair of diastereomers (Figure 2). Note that the cis diastereomers of each superwarfarin eluted ahead of the corresponding trans diastereomers during reversed phase chromatography (Table 1) as determined by Damin-Pernik et al. (8, 23), and confirmed by Huckle et al. (39), Kuijpers et al. (67), and Eadsforth et al. (37).

Figure 2.

Negative ion electrospray UHPLC-MS/MS SRM chromatograms of BDL, DFC, BDF, FCF, DFT, and BDF-d₄ (internal standard, 150 ng/mL) standards spiked into blank human plasma at 64 ng/mL.

During optimization of plasma extraction, a mixture of acetone and dichloromethane was found to extract not only lipophilic superwarfarins but also plasma lipids, as reported previously by Hauck et al. (58). These lipids contributed to the formation of emulsions during sample preparation and increased the frequency of reversed phase column clogging. Co-extraction of plasma lipids was avoided by precipitating plasma proteins with higher polarity solvents (acetonitrile:methanol; 90:10, v/v) while still recovering the superwarfarins.

All five superwarfarins formed abundant deprotonated molecules during negative ion electrospray which were used as precursor ions for SRM tandem mass spectrometry. Negative mode was used instead of positive mode due to lower background noise and lower limits of quantitation. Precursor ion, product ion and collision energy were optimized for each analyte (Table 1). The precursor ions for DFC, DFT and FCF used for SRM corresponded to the most abundant monoisotopic masses of the deprotonated molecules at m/z 443, m/z 537 and m/z 541, respectively. Due to lower background noise, the precursor ions used for SRM MS/MS of BDF, BDL and BDF-d₄ were m/z 523, m/z 527 and m/z 527, respectively, which corresponded to the heavy isotope of bromine of each deprotonated molecule.

Linearity and Lower Limit of Quantitation (LLOQ)

Calibration curves for cis- and trans-diastereomers of BDL, DFC, FCF, BDF, and DFT were linear across 3–4 orders of magnitude, ranging from 0.03 to 1305.18 ng/mL (Table 2). Log transformation of both concentration and peak area ratio provided excellent fit with coefficients of determination (R²) at >0.999 for each superwarfarin diastereomer (68). The LLOQ and LLOD values were experimentally determined for each diastereomer and ranged from 0.03 to 2.55 ng/mL and 0.02 to 1.28 ng/mL, respectively (Table 2).

Table 2.

Coefficient of determination (R²), calibration range, LLOQ, and LLOD for superwarfarin diastereomers in human plasma

Superwarfarin	Coefficient of determination, R2	Calibration range, ng/mL	LLOQ, ng/mL	LLOD, ng/mL
cis-BDL	0.9999	0.09–186.07	0.09	0.05
trans-BDL	0.9999	0.03–69.93	0.03	0.02
cis-DFC	0.9998	0.15–609.23	0.15	0.08
trans-DFC	0.9995	0.10–414.77	0.10	0.05
cis-FCF	0.9997	0.26–1074.29	0.26	0.13
trans-FCF	0.9997	0.24–973.71	0.24	0.12
cis-BDF	0.9998	0.19–793.74	0.19	0.10
trans-BDF	0.9999	0.31–1254.26	0.31	0.16
cis-DFT	0.9994	1.45–742.82	1.45	0.73
trans-DFT	0.9994	2.55–1305.18	2.55	1.28

Selectivity and Matrix Effect

UHPLC-MS/MS analysis of six different lots of blank human plasma showed no interfering signals (<10% of the analyte signal at the LLOQ) at the retention time of each pair of superwarfarin diastereomer (Figure 3). Matrix effects caused <10% deviation of analyte peak area for spiked blank matrix relative to spiked neat injection solvent at low and high superwarfarin concentrations. Therefore, the UHPLC-MS/MS method was found to be both highly selective and free from matrix effects.

Figure 3.

Selective detection of superwarfarins BDL, DFC, FCF, DFT, and BDF using UHPLC-MS/MS. Selected reaction monitoring (SRM) of blank human plasma (left column) is compared with blank plasma spiked with each superwarfarin at the LLOQ (right column).

Precision and Accuracy

Repeatability, intermediate precision and accuracy were evaluated at the LLOQ as well as at low, medium and high concentrations for each compound. Repeatability was evaluated using five replicates at the LLOQ, low, medium and high concentrations of all five superwarfarin rodenticides with RSDs of <12, <10, <7, and <8%, respectively (Table 3). The corresponding accuracy was 101–108, 100–109, 100–109, and 100–110%, respectively. Intermediate precision was evaluated using 15 replicates at the LLOQ, low, medium and high superwarfarin concentrations and was <11, <10, <8, and <9%, respectively, with accuracy of 100–107, 100–109, 101–105, and 100–106%, respectively (Table 3).

Table 3.

Precision, accuracy, and recovery of superwarfarin diastereomers measured using UHPLC-MS/MS

	Intra-day (n = 5)				Inter-day (n = 15)			(n = 5)
Superwarfarin	Spiked concn, ng/mL	Measured concn, ng/mL	Repeatability precision, RSD, %	Accuracy, %	Measured concn, ng/mL	Intermediate precision, RSD, %	Accuracy, %	Recovery, %
cis-BDL	0.05	0.05	9.1	102.1	0.05	10.4	103.0	86–96
	0.15	0.15	6.4	101.5	0.15	8.1	101.4
	29.65	31.50	6.9	106.2	30.99	6.6	104.5
	158.14	165.38	4.0	104.6	161.10	6.3	101.9
trans-BDL	0.01	0.01	8.2	104.0	0.01	10.5	106.5
	0.04	0.04	6.4	101.5	0.04	9.3	104.6
	7.85	7.96	6.4	101.4	7.661	7.5	102.4
	41.86	42.60	6.9	101.8	41.04	7.2	102.0
cis-DFC	0.14	0.14	10.0	102.5	0.13	8.6	105.1	87–95
	0.42	0.38	5.8	109.0	0.39	5.0	108.4
	84.47	82.32	6.6	102.5	85.49	5.4	101.2
	450.48	475.10	4.0	105.5	451.02	1.0	100.1
trans-DFC	0.11	0.12	11.1	107.5	0.11	9.5	104.8
	0.33	0.33	7.2	101.7	0.32	7.8	104.0
	65.53	68.10	3.3	103.9	66.28	5.0	101.1
	349.52	383.71	6.2	109.8	357.80	7.8	102.4
cis-FCF	0.29	0.28	5.4	101.7	0.28	6.3	103.7	88–103
	0.86	0.87	5.0	101.1	0.83	8.4	104.3
	172.39	178.95	3.0	103.8	177.67	4.4	103.1
	919.42	971.14	4.6	105.6	966.99	8.4	105.2
trans-FCF	0.21	0.22	9.0	104.6	0.22	9.4	103.8
	0.64	0.64	9.3	100.5	0.62	8.6	102.6
	127.61	134.19	3.4	105.2	131.47	6.4	103.0
	680.58	708.95	7.4	104.2	693.02	6.8	101.8
cis-BDF	0.22	0.22	7.6	101.2	0.22	8.1	100.2	81–98
	0.66	0.65	4.3	100.2	0.66	3.1	100.9
	131.01	132.16	5.7	100.9	134.52	6.6	102.7
	698.74	704.81	5.7	100.9	708.79	4.7	101.4
trans-BDF	0.28	0.29	5.3	103.1	0.29	5.2	102.6
	0.84	0.85	2.3	100.6	0.85	2.2	101.0
	168.99	167.87	5.3	100.7	166.86	6.5	101.3
	901.26	932.58	6.7	103.5	916.20	5.6	101.7
cis-DFT	1.59	1.64	4.8	103.4	1.60	5.2	100.5	68–94
	4.77	5.15	3.6	108.0	4.98	6.9	104.4
	119.23	128.86	5.1	108.1	121.38	7.9	101.8
	635.88	654.05	3.8	102.9	652.47	8.4	102.6
trans-DFT	2.41	2.31	7.6	104.0	2.27	8.1	105.9
	7.23	7.32	4.7	101.3	7.26	7.9	100.4
	180.77	181.14	2.0	100.2	177.38	5.2	101.9
	964.12	1007.16	3.3	104.5	1008.08	3.8	104.6

Recovery

Sample clean-up was carried out using protein precipitation (69). Methanol, acetonitrile, and solutions of 10–25% methanol in acetonitrile were compared, and acetonitrile containing 10% methanol provided both effective protein precipitation as well as analyte solubilization. Depending on initial concentration, the recoveries of BDL, DFC, FCF, BDF, and DFT ranged from 86 to 96%, 87 to 95%, 88 to 103%, 81 to 98%, and 68 to 94%, respectively (Table 3).

Stability

All five superwarfarins were stable in sample matrix on the laboratory bench at room temperature or in the autosampler at 4°C for 24 h (<10% loss compared with freshly prepared samples). Similar stabilities were observed for each analyte stored at –20 or –80°C for 60 days (Table 4).

Table 4.

Stability of superwarfarin diastereomers in extracted plasma during analysis (percentage remaining ± percentage CV)

Superwarfarin	Low concentration (n = 6)					High concentration (n = 6)
	Concn, ng/mL	–80°C (30 days)	–20°C (30 days)	4°C (24 h)	25°C (24 h)	Concn, ng/mL	–80°C (30 days)	–20°C (30 days)	4°C (24 h)	25°C (24 h)
cis-BDL	0.15	97.6 ± 6.7	98.9 ± 8.1	99.4 ± 7.5	96.9 ± 8.6	158.14	98.1 ± 4.7	99.8 ± 5.8	95.7 ± 5.5	93.3 ± 6.0
trans-BDL	0.04	98.0 ± 8.0	93.8 ± 5.1	94.8 ± 5.5	95.2 ± 9.5	41.86	93.9 ± 3.2	94.5 ± 2.8	97.2 ± 4.9	97.9 ± 5.8
cis-DFC	0.42	92.5 ± 5.1	93.0 ± 6.4	95.4 ± 9.3	94.5 ± 5.1	450.48	97.6 ± 4.3	95.2 ± 6.9	95.0 ± 5.0	93.8 ± 5.5
trans-DFC	0.33	93.5 ± 3.6	93.6 ± 3.8	93.3 ± 7.4	93.7 ± 3.6	349.52	95.2 ± 7.9	98.4 ± 6.5	95.8 ± 6.5	96.1 ± 8.0
cis-FCF	0.86	95.4 ± 7.9	90.6 ± 5.9	94.3 ± 5.6	91.0 ± 6.4	919.42	95.6 ± 6.0	94.6 ± 6.1	95.2 ± 7.2	94.9 ± 7.8
trans-FCF	0.64	96.1 ± 5.7	93.0 ± 9.5	93.3 ± 5.7	91.8 ± 8.1	680.58	96.6 ± 6.3	92.6 ± 6.3	90.7 ± 3.9	92.3 ± 6.7
cis-BDF	0.44	99.1 ± 8.4	98.8 ± 4.5	99.4 ± 4.4	97.9 ± 4.6	698.74	97.8 ± 2.3	97.7 ± 2.4	97.9 ± 2.7	99.0 ± 3.3
trans-BDF	0.84	96.5 ± 6.2	97.9 ± 6.6	98.7 ± 8.2	96.6 ± 5.9	901.26	97.3 ± 5.4	97.4 ± 5.9	94.6 ± 5.5	94.7 ± 5.9
cis-DFT	4.77	95.6 ± 5.6	97.4 ± 5.1	96.5 ± 7.8	96.7 ± 5.6	635.88	95.8 ± 6.3	94.0 ± 8.9	96.6 ± 6.7	93.8 ± 5.7
trans-DFT	7.23	97.8 ± 7.1	93.7 ± 6.8	99.1 ± 7.2	98.7 ± 4.7	964.12	98.9 ± 4.6	95.0 ± 5.2	97.1 ± 5.6	98.5 ± 4.4

Applications

Although all commercially available superwarfarins are synthesized as mixtures of stereoisomers, some synthetic reaction steps might show partial enantiomeric selectivity thereby altering the ratio of the resulting pairs of diastereomers. The synthetic routes for the commercially obtained superwarfarin standards are unknown, but their diastereomeric cis: trans ratios were determined. FCF showed the most equal cis: trans ratio with 52.4% cis and 47.6% trans, along with DFC which was 59.5% cis and 40.5% trans. The superwarfarins with the greatest differences in abundance between pairs of diastereomers were BDF-d₄ (81.8% cis and 18.2% trans), BDL (72.8% cis and 27.2% trans), DFT (36.3% cis and 63.7% trans) and unlabeled BDF (38.8% cis and 61.2% trans) (Figure 2). Note that the ratio of superwarfarin diastereomers in each standard reflects the synthetic pathway and purification procedures used in its preparation. Such differences are particularly evident for commercially purchased BDF and BDF-d₄. This divergence can also be seen between commercial BDF and DFC and the products of a seven-step synthesis used by Heerden et al. (19, 26), who reported a cis: trans ratio for BDF of 44% cis and 56% trans and a cis: trans ratio for DFC of 40% cis and 60% trans. The abundance of cis-BDF is lower in the commercial standard and for DFC the cis: trans ratio is reversed.

Superwarfarin rodenticides are registered for use in many countries with some regional differences (70). For example, the United Kingdom restricts DFT to indoor use (71), and the United States permits the use of BDF, DFC, BDL and DFT but not FCF (72). Because superwarfarins have been used in terrorist attacks, added to food and pharmaceutical products, and even mixed with narcotics (73), the ratio of superwarfarin diastereomers, determined using this method of separation, might be useful to identify sources of superwarfarins during forensic investigations (24).

During 2018, an outbreak of coagulopathy occurred in Illinois that was attributed to smoking a synthetic cannabinoid product contaminated with superwarfarin rodenticides (64, 65). We used this UHPLC-MS/MS method to analyze plasma from 31 patients (21 men and 10 women) treated during this outbreak, and recently published an observational clinical case report regarding one of these poisoning victims (66). Among these 31 patients, plasma from two patients contained only BDF, plasma from 11 patients contained both BDF and DFC, and plasma from 18 patients contained BDF, DFC, and BDL. Superwarfarins FCF and DFT were not detected in any plasma specimens. Upon hospitalization, the patient plasma concentration of BDF, DFC, and BDL ranged from 2.7 to 1861.8 ng/mL, 0.6 to 41.4 ng/mL, and 0.1 to 138.8 ng/mL, respectively [for clinical analysis of these data, see (74)]. As an example, Figure 4 shows UHPLC-MS/MS SRM chromatograms for measurement of BDL, DFC and BDF in the plasma from one of these poisoning victims, corresponding to plasma concentrations of 52.6 ng/mL, 30.6 ng/mL, and 1861.8 ng/mL, respectively.

Figure 4.

UHPLC-MS/MS SRM chromatograms of plasma from a patient who inhaled synthetic cannabinoid products tainted with superwarfarin rodenticides. The plasma concentrations of BDL, DFC, and BDF at the time of hospital admission were 52.6 ng/mL (54.4% cis-BDL), 30.6 ng/mL (98.5% cis-DFC) and 1861.8 ng/mL (39.5% cis-BDL), respectively. Note the absence of trans-DFC, indicating a large difference in half-life between cis- and trans-DFC.

Since anticoagulant rodenticide products contain a single superwarfarin, the presence of multiple superwarfarins in the plasma of a single patient suggests deliberate adulteration rather than accidental contamination of the ingested synthetic cannabinoids. One non-malicious explanation that might explain the presence of multiple superwarfarins is a mistaken belief that they potentiate the effect of synthetic cannabinoids (75).

Although cis-diastereomers of DFC were present in plasma from 29 patients, no trans-diastereomers of DFC were detected in any specimens (Figure 4). This observation is consistent with previous reports that trans-DFC is eliminated more rapidly than the cis form (23, 43). However, the ratio of cis- and trans-diastereomers of both BDL and BDF remained nearly constant in plasma obtained for up to 28 days, suggesting that both pairs of diastereomers were eliminated at similar rates.

Current treatment options for superwarfarin poisoning are limited to replenishing VK₁ levels and activated clotting factors. The disadvantage of these treatments is that toxic levels of superwarfarins can remain in a patient for up to a year, due to the 20–30-day half-lives. This can result in prolonged and expensive supplementation with oral VK₁ (64). On-going research supported by this analytical method is focused on hastening the elimination of superwarfarins from poisoning victims.

Conclusions

A fast and sensitive analytical method based on a 10-min UHPLC-MS/MS assay was developed and validated for the separation and quantitative analysis of the pairs of diastereomers of BDL, DFC, FCF, BDF, and DFT in human plasma. Analysis of plasma specimens from coagulopathy patients using this method facilitated the identification and quantification of the superwarfarin causative agents and provided evidence of asymmetric elimination of DFC and BDL diastereomers.