Deploying Validated Mass Spectrometry for Frontline Detection and Treatment of Human Poisoning by Long-Acting Anticoagulant Rodenticides

Abstract

Derived from the same natural anticoagulant as warfarin (dicoumarol), long-acting anticoagulant rodenticides (LAARs) or superwarfarins have much longer half-lives in human blood than warfarin (weeks instead of hours) and are more potent inhibitors of the same enzyme, vitamin K epoxide reductase component 1. While used effectively worldwide as rodenticides, LAARs can elicit severe, protracted, life-threatening coagulopathy in humans at blood concentrations >10 ng/mL leading to numerous accidental and intentional poisonings annually. To facilitate timely identification and quantitative analysis of LAARs in patients presenting unexplained severe, protracted, life-threatening coagulopathy, several analytical methods have been developed, all of which are based on electrospray liquid chromatography–mass spectrometry (LC–MS). In this perspective, we evaluated and compared these LC–MS methods in terms of validation, simultaneous detection of multiple LAARs, measurement of individual stereoisomers, and clinical applications.

Graphical Abstract

BACKGROUND

Formed by the fungus Melilotus officinalis (L.) Pall acting on the plant natural product coumarin, bis-hydroxycoumarin (dicoumarol) (Figure 1) was identified in 1940 as a cause of bleeding disorders in cattle consuming sweet clover hay.¹ Dicoumarol acts indirectly as an anticoagulant by inhibiting vitamin K epoxide reductase component 1 (VKORC1) which converts vitamin K epoxide to vitamin K hydroquinone. Vitamin K hydroquinone is a cofactor required for the action of γ-glutamyl carboxylase which carboxylates and activates specific proteins in the coagulation cascade (factors II, VII, IX, and X).²

Figure 1.

Chemical structures of dicoumarol, warfarin, some other first generation anticoagulant rodenticides, and examples of second generation long-acting anticoagulant rodenticides (LAARs).

VKORC1 is highly conserved among vertebrates,³ and although dicoumarol also inhibits VKORC1 in rats and mice, it acts too slowly to be effective as a rodenticide. To obtain more potent inhibitors of VKORC1, analogues of dicoumarol were synthesized for testing, which resulted in the discovery of warfarin (4-hydroxy-3-(3-oxo-1-phenylbutyl) coumarin) and other first generation anticoagulant rodenticides including coumachlor, coumafuryl, and coumatetralyl,⁴ and the structurally related indandione anticoagulants such as diphenadione (diphacinone) (Figure 1).⁵ Although warfarin is currently used in human medicine as the anticoagulant drug Coumadin,⁶ mutations in VKORC1 among rodents resulted in the development of resistance to warfarin and other first generation rodenticides that was recognized as early as 1960.⁷

To overcome rodent resistance to first generation anticoagulants, more potent and longer acting second generation dicoumarol analogues were developed, which are known as superwarfarins or long-acting anticoagulant rodenticides (LAARs) and include bromadiolone, brodifacoum, difenacoum, difethialone, and flocoumafen (Figure 1). Based on the 4-hydroxycoumarin moiety, LAARs contain bulky aromatic substituents at the carbon-3 position which improve binding to and inhibition of VKORC1,⁸ enhance hydrophobicity and accumulation, reduce metabolic transformation and elimination, and contribute to long half-lives.^9,10 Due to conservation of VKORC1 among vertebrates, LAARs are also potent and long-lasting anticoagulants in humans. For example, the half-lives of LAARs in humans are 16–62 days¹¹ compared with 35 h for warfarin.¹²

The extensive use of LAARs results in thousands of accidental and intentional human poisonings worldwide each year. For example, there were 3,342 cases of human poisonings reported in the United States in 2020¹³ and 3,586 cases in 2019.¹⁴ Although most LAAR poisonings are accidental, approximately 7% of these cases are intentional and involve suicide, homicide, or other malicious acts.^13-15 Clinical symptoms of severe coagulopathy typically occur within 24–72 h after exposure to LAARs, and hospital laboratory tests characteristically show prolonged prothrombin times and international normalized ratio (INR) values that can persist for weeks to months. According to the Centers for Disease Control,¹⁶ confirmation that one or more LAARs caused a particular case of severe coagulopathy requires either knowledge of exposure, which can be confounded by the delay in developing clinical symptoms, or by identification of the causative agent in whole blood or plasma using appropriate laboratory tests.

Once identified, LAAR poisoning can be treated effectively, thereby minimizing mortality. The current standard of care is high dose intravenous or oral vitamin K₁ followed by daily high dose oral vitamin K₁ for several weeks or months to restore and maintain normal coagulation until LAAR plasma concentration drops to ≤10 ng/mL.^17,18 Even after normalization of blood coagulation rates, premature termination of vitamin K₁ therapy can result in rebound toxicity if LAAR levels remain above safe limits, which is a particular problem given that vitamin K₁ might be required for several months increasing the risk of patient noncompliance.¹⁹ Therefore, quantitative analysis of LAARs in blood using an appropriate laboratory method is critical to determining when it is safe to terminate vitamin K₁ therapy. This perspective describes and compares the analytical methods that have been developed for the identification and quantitative analysis of LAARs in human blood and plasma.

METHODS FOR IDENTIFICATION AND QUANTITATIVE ANALYSIS OF LAARS IN HUMAN BLOOD AND PLASMA

A variety of methods have been reported for the identification and quantitative analysis of LAARs in rodenticide products and animal blood, plasma, and tissues that include high-performance liquid chromatography (HPLC) with ultraviolet absorbance,²⁰ fluorescence,²¹ electrochemical,²² or mass spectrometric (MS) detection.²³ Fewer methods have been reported for the measurement of one or more LAARs in human blood or plasma (Table 1). Note that serum is not always suitable for measuring exposure to LAARs because the preparation of serum requires that clotting take place before the removal of precipitated proteins and LAARs inhibit blood clotting. Note also that urine is not an appropriate matrix for measuring exposure to LAARs as these anticoagulants are not excreted in urine.^24,25

Table 1.

Electrospray LC–MS Methods for Measuring LAARs in Human Whole Blood or Plasma

Ref	Sample preparation	LAAR recovery %	LC column	Resolution of stereoisomers	MS type	Validated calibration range ng/mL
Grobosh et al., 2006.35	LLEa	-	HPLC	No	Quadrupole	Five LAARs
	Chlorobutane		C18		MS	80–120
Jin et al., 2007.30	LLE	82–85	HPLC	No	Ion trap	Bromadiolone
	Ethyl acetate		C18		MS/MS	0.5–100
Vindenes et al., 2008.29	Acetonitrile protein precipitation	-	HPLC	No	Hexapole	Bromadiolone
			Phenyl		MS	10–1000
Lo et al., 2008.17	LLE dichloromethane/dichloroethane	-	HPLC	No	Triple quadrupole	-
			C8		MS/MS
Yan et al., 2012.42	LLE	54–77	HPLC	No	Triple quadrupole	Bromadiolone, brodifacoum
	Ethyl acetate		C18		MS/MS	0.5–100
Schaff and Montgomery 2023.27	Acetonitrile protein precipitation and SPEb	58–94	HPLC	No	Orbitrap	-
			C18		MS/MS
Bidny et al., 2015.38	LLE	72–93	UHPLC	No	Triple quadrupole	Five LAARS
	Ethyl acetate		C18		MS/MS	2–200
Gao et al., 2018.39	SLEc	60–110	UHPLC	No	Orbitrap	Five LAARs
	Ethyl acetate		C18		MS/MS	0.07–200
Guo et al., 2018.43	Acetonitrile protein precipitation and SPE	81–113	UHPLC	No	Triple quadrupole	Three LAARs
			C18		MS/MS	1–2000
Qiao et al., 2018.37	LLE	50–63	HPLC	No	Triple quadrupole	Four LAARs
	Ethyl acetate		C18		MS/MS	0.5–50
Nosal et al., 2020.33	Acetonitrile/methanol protein precipitation	68–103	UHPLC	Pairs of diastereomers	Triple quadrupole	Five LAARs
			C18		MS/MS	0.013–2.41
Arens et al., 2021.36	Methanol protein precipitation	41–93	HPLC	No	Triple quadrupole	Three LAARs
			C18		MS/MS	1–100
Nosal et al., 2021.28	Acetonitrile/methanol protein precipitation	81–98	HPLC	All stereoisomers	Triple quadrupole	Three LAARs
			Chiral		MS/MS	0.87–652
Yu et al., 2024.44	LLE	-	HPLC	No	Triple quadrupole	Brodifacoum-
	Ethyl acetate		C18		MS/MS

^aLLE = liquid/liquid extraction.

^bSPE = solid phase extraction.

^cSLE = supported liquid extraction

Gas chromatography–MS is not ideal for the analysis of LAARs because these compounds undergo pyrolysis in the inlet of the gas chromatograph.²⁶ Instead, all of the published methods for measuring LAARs in human blood or plasma utilize HPLC or ultrahigh pressure liquid chromatography (UHPLC) interfaced with MS (Table 1). Although both electrospray and atmospheric pressure chemical ionization are compatible with HPLC and UHPLC, atmospheric pressure chemical ionization was found to be less sensitive than electrospray for the measurement of LAARs and has therefore not been used.²³ All of the published methods for measuring LAARs in human specimens have utilized negative ion electrospray, except for one which used positive ion electrospray and was only validated for the identification but not quantitative analysis of LAARs in human blood.²⁷ In that study, Schaff and Montgomery²⁷ found that most LAARs can be measured as protonated molecules except for bromadiolone, which eliminated water in the atmospheric chemical ionization source and was measured instead as the [MH-H₂O]⁺ ion. Negative ion electrospray MS is therefore preferred to positive ion mode since all LAARs have been found to form abundant deprotonated molecules.

CHROMATOGRAPHIC SEPARATIONS OF LAARS DURING LC–MS

To date, all the chromatographic methods used for the LC–MS analysis of LAARs in human blood and plasma have utilized reversed phase chromatography except for that of Nosal et al.,²⁸ who used a chiral column. Most of the reversed phase columns used in these methods contained a C₁₈ stationary phase, except for the methods reported by Vindenes et al.,²⁹ who used a phenyl column; and Lo et al.,¹⁵ who used a C₈ column (Table 1). All of the chromatographic methods separating multiple LAARs and even some methods reported for single LAAR measurement utilized solvent gradients, except for that of Jin et al.,³⁰ who measured bromadiolone using an isocratic mobile phase. Gradient elution enables faster separation of multiple LAARs while improving the peak shape and concentrating each analyte as it elutes from the column, which improves the limit of quantitation during MS detection. Blood and plasma are complex matrices, and gradient elution also helps regenerate the column in preparation for the next analysis. The use of UHPLC in place of HPLC can also shorten retention times, improve peak shape, and further enhance MS detection. Four LC–MS methods have used UHPLC with C₁₈ stationary phases (Table 1).

All LAARs contain two chiral centers and are synthesized commercially as mixtures of four stereoisomers (Figure 1).²³ Although only chiral columns can separate all four of these stereoisomers, pairs of diastereomers can be separated by using reversed phase chromatography.^31,32 However, only Nosal et al.³³ reported the separation, identification, and quantitative analysis of diastereomeric pairs of five LAARs extracted from human plasma (Figure 2A). The key difference between the method of Nosal et al.³³ and most other reversed phase separations used for LAARs is the use of acetonitrile instead of methanol as the organic modifier. As observed previously by Hauck et al.,²³ the use of methanol for reversed phase chromatography of LAARs typically results in a single peak of unresolved stereoisomers for each LAAR.

Figure 2.

A) Simultaneous UHPLC–MS/MS analysis of LAARs showing separation of pairs of diastereomers. Selected reaction monitoring (SRM) MS/MS was used with collision-induced dissociation on a triple quadrupole mass spectrometer (Used with permission from Nosal et al.³³); and B) chiral LC–MS/MS on a triple quadrupole mass spectrometer of LAARs showing separation of all four stereoisomers (used with permission from Nosal et al.²⁸).

In the only LC–MS method utilizing a chiral column, Nosal et al.²⁸ reported chiral separations of all four stereoisomers of each of three LAARs extracted from human plasma (Table 1 and Figure 2B). The chiral column used in this study consisted of quinine immobilized on superficially porous silica, and the mobile phase was methanol/water with a linear gradient of increasing formic acid. This mixed-mode chiral column enabled separation based primarily on hydrogen bonding and π–π stacking interactions between immobilized chiral quinine and the LAAR stereoisomers.²⁸

Commercial LAARs contain an unequal distribution of stereoisomers as a result of isotope effects during synthesis or use of starting materials containing unequal amounts of stereoisomers.²⁸ Therefore, the ability to measure the ratios of diastereomers and enantiomers of LAARs in poisoning cases might be useful during forensic investigations by facilitating identification of the source of the material. Also, the stereoisomers of LAARs in human plasma have been found to be eliminated at different rates, which might aid in the development of new LAARs that have shorter half-lives in humans.^28,33,34 A potential disadvantage of the separation of LAARs extracted from human plasma into the corresponding stereoisomers might be loss of sensitivity compared with LC–MS methods that measure just one peak containing all four stereoisomers for each LAAR. However, the lower limit of quantitation (as indicated by the validated calibration ranges in Table 1) for the chiral LC–MS method of Nosal et al.²⁸ is among the best at 0.87 ng/mL LAAR in human plasma, and the method of Nosal et al.³³ reporting separation of diastereomeric pairs of LAARs is the most sensitive of all at 0.013 ng/mL. Another disadvantage of separating stereoisomers of multiple LAARs during the same analysis is that more time is required than when separating LAARs into just one chromatographic peak each. For example, the chiral LC–MS separation of Nosal et al.²⁸ required 12 min to resolve all 4 isomers of 3 LAARs, whereas 10 min were required to resolve pairs of diastereomers of 5 LAARs,³³ and as little as 5 min were needed to resolve 5 LAARs into a single peak each.³⁵

SAMPLE PREPARATION AND METHOD VALIDATION

A variety of sample preparation methods have been used for the measurement of LAARs in human blood and plasma that include the precipitation of plasma proteins using acetonitrile or methanol, liquid/liquid extraction with organic solvents, and combinations of plasma protein precipitation followed by solid phase extraction (Table 1). Based on LAAR recoveries, acetonitrile precipitation (58–94%²⁷) of protein was modestly more efficient than methanol precipitation (41–91%³⁶), and a combination of acetonitrile and methanol showed the best recovery (81–98%²⁸). Liquid/liquid extraction using ethyl acetate was used by five methods but generally showed lower recoveries (from 50 to 63%³⁷ to 72–93%³⁸) than did protein precipitation using acetonitrile or acetonitrile/methanol (Table 1).

MASS SPECTROMETRY OF LAARS AND METHOD VALIDATION

All of the methods reported for the analysis of LAARs in human blood and plasma identify LAARs by comparing LC–MS chromatographic retention times and molecular masses to those of the authentic standards. A variety of mass spectrometers have been used including single quadrupole, triple quadrupole, ion trap, and Orbitrap instruments (Table 1). Some methods utilized high-resolution MS with accurate mass measurement on Orbitrap mass spectrometers,^27,39 but most methods used tandem mass spectrometry (MS/MS), which enabled not only high selectivity but also high sensitivity. However, not all of these methods are suitable for the qualitative and quantitative analysis of all LAARs in human blood or plasma.

Quantitative methods should be validated according to standards set by organizations such as the U.S. Food and Drug Administration⁴⁰ and the European Medicines Agency⁴¹ and should include data such as calibration curve and range, recovery concentrations following sample preparation, assessment of method accuracy and precision, and analyte stability. Although qualitative methods intended solely for LAAR identification would not require assessment of all these parameters (such as calibration curve and range), they should include measures such as accuracy to prevent false positive results and recovery and limit of detection to prevent false negative results.

Although all 14 methods described sample preparation procedures, four did not report any recovery data; and two did not provide any calibration range data, which makes them unfit for qualitative or quantitative analysis of LAARs (Table 1). Schaff and Montgomery²⁷ provided recovery data but not calibration range making the method suitable for its intended purpose of identifying LAARs in blood. The remaining nine methods provided some method validation data such as recovery and calibration range and had lower limits of quantitation below 10 ng/mL, which is the threshold above which LAARs poisoning becomes clinically relevant (Table 1).

CLINICAL APPLICATIONS OF LC–MS TO LAARS POISONING

The clinical utility of most of the validated LC–MS methods has been demonstrated by identifying and measuring LAARs as the causative agents of unexplained severe, protracted, and life-threatening coagulopathy in humans. For example, Jin et al.,³⁰ Yan et al.,⁴² Guo et al.,⁴³ and Qiao et al.³⁷ used their LC–MS methods to identify bromadiolone and measure its plasma concentrations for several clinical cases. Yan et al.,⁴² Gao et al.,³⁹ Guo et al.,⁴³ and Bidny et al.³⁸ identified and measured brodifacoum in plasma from several patients, and Gao et al.³⁹ identified and measured both bromadiolone and brodifacoum in blood from a cadaver following suicide. Finally, Nosal et al.^28,33 identified and measured both pairs of diastereomers for five LAARs as well as all four stereoisomers of bromadiolone, brodifacoum, and difenacoum in plasma from 31 patients (21 men and 10 women) who were treated in 2018 during a severe, life-threatening coagulopathy outbreak in Illinois, USA.

CONCLUSIONS AND FUTURE PERSPECTIVES

Although LAARs have been measured using a variety of analytical methods, only electrospray LC–MS has emerged as being fit for the purpose of measuring multiple LAARs in human blood and plasma. To provide reproducible results for clinical applications, analytical assays must be validated according to established guidelines. Out of 14 LC–MS methods that have been reported for measuring LAARs in human blood and plasma, 10 were validated for LAAR identification, and 9 were validated for both qualitative and quantitative determination of LAARs. Among these, only three methods measured all five commercially used LAARs (Bidny et al.,³⁸ Gao et al.,³⁹ and Nosal et al.²⁸), whereas the others measured 4, 3, 2, or just 1 LAAR. Now that the utility of using validated LC–MS methods to identify and measure blood and plasma concentrations of LAARs has been demonstrated for clinical and forensic specimens, qualitative and quantitative LC–MS analysis will hopefully become more widely used for the diagnosis and long-term management of unexplained severe, protracted, life-threatening coagulopathy in humans. Whether substantial cost savings could be realized by deploying these user-friendly, robust analytical methods to frontline healthcare providers who treat LAAR-poisoned people remains to be determined in prospective studies.

ABBREVIATIONS

LAARs

long-acting anticoagulant rodenticides

LC–MS

liquid chromatography–mass spectrometry

VKORC1

vitamin K epoxide reductase component 1

HPLC

high-performance liquid chromatography

MS

mass spectrometry

UHPLC

ultrahigh pressure liquid chromatography

LLE

liquid/liquid extraction

SPE

solid phase extraction

SLE

supported liquid extraction

SRM

selected reaction monitoring

Biographies

Richard B. van Breemen is Professor of Pharmaceutical Sciences in the College of Pharmacy, the Global Hemp Innovation Center, and the Linus Pauling Institute at Oregon State University. He received his B.A. in Chemistry from Oberlin College and Ph.D. in Pharmacology and Experimental Therapeutics from the Johns Hopkins University. After postdoctoral training in mass spectrometry at Johns Hopkins, he taught at North Carolina State University and the University of Illinois at Chicago. In addition to LAARs, his research concerns the discovery of pharmacologically active natural products using affinity selection-mass spectrometry, and the safety and efficacy of botanical dietary supplements.

Bianca Flores is a graduate student at Oregon State University in the Pharmaceutical Sciences Ph.D. program. She earned an A.S. degree in Veterinary Technology, a B.A. in Chemistry, and a minor in Spanish from California Lutheran University. She also completed a year-long research experience in the PREP postbaccalaureate program at the University of Rochester School of Medicine and Dentistry with Dr. Megan Falsetta. Her research concerns the use of bile acid sequestrants to treat superwarfarin poisoning and prevent long-term adverse effects.

Israel Rubinstein, M.D. is a practicing pulmonologist and Professor of Medicine at the University of Illinois College of Medicine in Chicago. He develops and tests novel, safe, and efficacious countermeasures to treat poisoned patients. Dr. Rubinstein has authored more than 250 peer-reviewed scientific papers and holds several issued and pending US and international patents in his field of research. He is also Clinical and Reviews Editor, Nanomedicine: Nanotechnology, Biology, and Medicine. He cofounded ResQ Pharma, Inc, and EnSol Therapeutics, LLC, and was Board Member and Director of Advanced Life Sciences, a publicly traded, clinical stage, biopharmaceutical company in Woodridge, Illinois.

Douglas L. Feinstein is Research Professor at the University of Illinois at Chicago and Senior Research Career Scientist at the Jesse Brown VA Medical Center. He received his B.S. from MIT, Ph.D. from the Johns Hopkins University, and postdoctoral training at the Scripps Clinic in San Diego, the University of Uppsala in Sweden, and the University of Lausanne in Switzerland. He was a faculty member at Cornell University Medical College in New York before moving to Chicago. Dr Feinstein’s research concerns developing treatments for the neuroinflammation induced neuropathology in MS and AD and counteracting the neurotoxicity of long-acting anticoagulant rodenticides.