Detecting and quantifying marijuana metabolites in serum and urine of 19 dogs affected by marijuana toxicity

Abstract

Veterinarians diagnose marijuana toxicity based on clinical signs and history, or in conjunction with an over-the-counter (OTC) human urine drug screen. With the legalization of recreational marijuana use becoming more prevalent in the United States, a more accurate test to aid in the diagnosis of canine marijuana toxicity is needed. We collected urine and serum samples from 19 dogs with confirmed or suspected marijuana toxicosis from multiple veterinary hospitals and analyzed them with a novel UPLC-MS/MS method. Calibrations from 0.1 to 100 ng/mL and QC materials were prepared. Samples were extracted, purified, and eluted with solid-phase extraction. Urine samples were tested with an OTC human urine drug screen. The limit of detection (LOD) and lower limit of quantification (LLOQ) ranges for marijuana metabolites in serum were 0.05–0.25 ng/mL and 0.1–0.5 ng/mL, respectively. In urine, the LOD and LLOQ ranges for the metabolites were 0.05–0.1 ng/mL and 0.1–0.5 ng/mL, respectively. In serum, median and range of metabolite concentrations (ng/mL) detected included: THC, 65.0 (0.14–160); 11-OH-Δ⁹-THC, 4.78 (1.15–17.8); 11-nor-9-carboxy-Δ⁹-THC, 2.18 (0.71–7.79); CBD, 0.28 (0.11–82.5); and THC-glucuronide, 2.05 (0.72–18.3). In the 19 urine samples, metabolite: creatinine (ng: mg) values detected included: THC, 0.22 (0.05–0.74); 11-OH-Δ⁹-THC, 0; 11-nor-9-carboxy-Δ⁹-THC, 1.32 (0.16–11.2); CBD, 0.19 (0.12–0.26); THC-COOH-glucuronide, 0.08 (0.04–0.11); and THC-glucuronide, 0.98 (0.25–10.7). Twenty of 21 urine samples tested negative for THC on the urine drug screen. All 19 serum samples contained quantifiable concentrations of THC using our novel UPLC-MS/MS method. Utilizing a UPLC-MS/MS method can be a useful aid in the diagnosis of marijuana toxicosis in dogs, whereas using an OTC human urine drug test is not a useful test for confirming marijuana exposure in dogs because of the low concentration of THC-COOH in urine.

Marijuana (Cannabis sativa) is a plant that has been used medicinally and recreationally for centuries.^3,5 Tetrahydrocannabinol (THC) is the main constituent that is responsible for the psychoactive effects, and cannabidiol (CBD) is the primary non-psychoactive metabolite.^3,5 Thirty percent of the states in the United States have legalized recreational marijuana, and this could lead to more canine exposures with the increasing availability of the drug to the public.⁴ Dogs are most commonly exposed to marijuana via ingestion of the plant material or edible marijuana products.^3,14 Most intoxications are not lethal but can result in severe clinical signs that require patient care.^9,12 One retrospective study evaluated a cohort of 125 dogs in Colorado and reported a positive correlation between the number of marijuana toxicoses in dogs and the number of medical marijuana licenses issued from 2005 to 2010.¹² Additionally, the concentration of THC in cannabis samples seized by the Drug Enforcement Administration from 1995 to 2018 has more than tripled,¹¹ which may explain the high morbidity associated with marijuana exposure in dogs, prompting their owners to seek medical attention for their pets.

Common clinical signs of marijuana intoxication in dogs include ataxia, urine dribbling, depression, mydriasis, and hyperesthesia.^3,5,9,12,14 The nonspecific signs of marijuana toxicity pose a diagnostic challenge to the clinician if the owner is unaware or reluctant to discuss exposure.¹⁶ Confirmatory tests are largely unavailable for suspected marijuana intoxication in dogs. Some clinicians utilize an over-the-counter (OTC) human urine drug screen that most commonly measures 11-nor-9-carboxy-Δ⁹-tetrahydrocannabinol (THC-COOH); however, the canine marijuana metabolite profile is unknown.¹⁶

We quantified marijuana metabolites in serum and urine of dogs that had been presented because of suspected marijuana toxicity. The main human marijuana metabolites are 11-hydroxy-Δ⁹-tetrahydrocannabinol (11-OH-THC), THC-COOH, and CBD.⁸ Our primary objective was to detect and quantify marijuana compounds in otherwise healthy dogs exposed to marijuana by developing and validating an ultra-performance liquid chromatography–tandem mass spectrometer (UPLC-MS/MS) method. Our secondary objective was to evaluate the usefulness of a human urine drug screen in detecting marijuana and its related products in otherwise healthy but exposed dogs. Our hypothesis was that, upon marijuana exposure, dogs produce low or negligible amounts of THC-COOH in urine compared to other metabolites, leading to a high proportion of false-negative tests when using the OTC human urine drug test.

We enrolled in our pilot study otherwise healthy dogs that had been presented with known (owner-observed exposure) or suspected (based on clinical signs and history) marijuana ingestion to the Kansas State University Veterinary Health Center (KSU-VHC; Manhattan, KS, USA), University of Illinois Veterinary Teaching Hospital (UI-VTH; Urbana, IL, USA), and 2 private practices (Animal ER of University Park, Bradenton, FL, USA [Animal ER] and Veterinary Emergency + Critical Care, Las Vegas, NV, USA [VECC]). Institutional Animal Care and Use Committee (IACUC) approval was obtained at the 2 universities, and client consent was obtained from all dog owners prior to enrollment. We included 21 dogs in our study: 19 dogs with clinical signs of marijuana toxicity and 2 client-owned dogs with no known exposure to marijuana, which were included to verify that there were no endogenous metabolites that could be misidentified as metabolites of interest. We collected ~3 mL of blood from a peripheral vein and 5 mL of urine by cystocentesis or voiding from each dog. Serum and urine samples were shipped overnight on ice to the Kansas State Veterinary Diagnostic Laboratory and were stored at −80°C for up to 6 mo until analysis.^7,13,15 Urine marijuana metabolites were standardized to urine creatinine for each patient. Urine creatinine was measured via the Jaffe method (Cobas c501; Roche). Urine samples, at room temperature, were tested by an OTC human urine drug screen (iScreen 2 panel drug test THC/coc; AlcoPro) to detect THC (THC-COOH target analyte).²

The following standards were utilized: THC, THC-d₃, 11-OH-THC, 11-OH-THC-d₃, THC-COOH, THC-COOH-d₃, CBD, CBD-d₃, THC-COOH-glu, THC-COOH-glu-d₃ (all from MilliporeSigma), and THC-glu (ElSohly Laboratories). Sera and urine samples, calibrators, blanks (negative dog serum or urine), and quality controls (QC) were extracted and cleaned-up by solid-phase extraction (SPE; Oasis HLB Prime µElution 96-well plates; Waters). For serum samples, 100 µL of each calibration working solution, QC solution, and the serum sample were added to a 1.5-mL centrifuge tube, respectively. A 50-µL aliquot of the internal standard (IS) solution was placed into each centrifuge tube along with 150 µL of 0.1% formic acid in acetonitrile, vortexed for 10 s, and centrifuged for 10 min at 13,000 × g at 4°C. For urine samples, 100 µL of each calibrator working solution, QC solution, and the urine sample were added into a 48-well non–tissue culture-treated plate. Fifty µL of the IS solution was placed into each well along with 150 µL of 0.1% formic acid in acetonitrile. Subsequently, the plate was shaken for 20 min and centrifuged at 2,000 × g for 30 min at room temperature. The supernatant was transferred to a new centrifuge tube and diluted with 400 µL of water. The diluted sample was placed into a 96-well plate. Following a wash with 2 × 250 µL of 25% methanol, the components were eluted with 50 µL of 90/10 acetonitrile/methanol. Fifty μL of 0.1% formic acid were added onto each well of the 96-well collection plates before the UPLC-MS/MS analyses.

Concentrations of the marijuana metabolites in serum and urine samples were determined (Xevo TQ-S triple quadrupole mass spectrometer; Waters; Fig. 1, Suppl. Fig 1). Briefly, an Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 50 mm) was held at 45°C with eluents composed of mobile phase A (0.1% formic acid in water) and mobile phase B (0.1% formic acid in acetonitrile). The flow rate was 0.4 mL/min. The following gradient program was used: from 0 to 3.5 min phase A (25–5%), at 3.5 min phase A (5%), hold for 0.5 min, and at 4.2 min phase A (25% for 0.8 min). A positive ionization mode was used (+ESI).^1,8 The ions were monitored using multiple reaction monitoring (MRM) mode (Table 1).

Figure 1.

Multiple reaction monitoring (MRM) chromatograms from analysis of a negative control dog serum. A, C, E, G, I, and K in the left panel are the MRM chromatograms of negative canine serum. B, D, F, H, J, and L in the right panel are the MRM chromatograms of cannabinoids (CBD, THC, 11-OH-THC, THC-COOH, THC-glu, THC-COOH-glu, respectively) spiked at 10 ng/mL in negative canine serum.

Table 1.

Multiple reaction monitoring method development transition ions for each cannabinoid in canine serum and urine.

Metabolite	Qualifier ion	Quantifier ion
CBD	315.237 → 135.180	315.237 → 193.139
CBD-d3	318.257 → 138.217	318.257 → 196.176
THC	315.252 → 135.167	315.252 → 192.992
THC-d3	318.258 → 138.168	318.258 → 195.993
11-OH-THC	331.253 → 193.060	331.258 → 313.138
11-OH-THC-d3	334.258 → 196.060	334.258 → 316.138
THC-COOH	345.223 → 192.995	345.223 → 299.087
THC-COOH-d3	348.223 → 302.082	348.223 → 195.990
THC-glu	491.470 → 177.200	491.470 → 315.300
THC-COOH-glu	521.474 → 301.300	521.474 → 345.363
THC-COOH-glu-d3	524.494 → 304.169	524.494 → 348.234

The developed method was validated in reference partly to the FDA guidelines for the Bioanalytical Method Validation Guidance for Industry, with precision ≤20% and accuracy of 80–120%.¹⁷ A signal-to-noise ratio of 3:1 was generally considered acceptable for estimating the limit of detection (LOD). A typical signal-to-noise ratio of 10:1 of the analyte in the negative control sample was considered acceptable for estimating the lower limit of quantification (LLOQ). For serum samples, the calibration curve was linear from 0.1 (CBD and THC) or 0.5 (11-OH-THC, THC-COOH, and THC-glu) to 100 ng/mL with a correlation coefficient (R²) of at least 0.99. For urine samples, the calibration curve was linear from 0.1 (CBD and THC), 0.25 (11-OH-THC and THC-COOH) or 0.5 (THC-glu and THC-COOH-glu) to 100 ng/mL with R² of at least 0.99 (Table 2). For serum, deviations in QC samples (n = 9) were <11.4% of nominal concentrations for all of the compounds. The accuracy of the assay for different analytes was 93.4–112% of their nominal values for calibrators. For urine, deviations in QC samples (n = 9) were <16.8% of nominal concentrations for all of the compounds. The accuracy of the assay for different analytes was 84.0–110% of their nominal values for calibrators.

Table 2.

Limit of detection (LOD) and lower limit of quantification (LLOQ) for cannabinoids in canine serum and urine (ng/mL).

	THC	11-OH-THC	THC-COOH	CBD	THC-glu	THC-COOH-glu
Serum
LOD	0.05	0.10	0.10	0.05	0.25
LLOQ	0.10	0.5	0.5	0.10	0.50
Urine
LOD	0.05	0.10	0.10	0.05	0.10	0.10
LLOQ	0.10	0.25	0.25	0.10	0.50	0.50

The 21 canine serum/urine pairs included: 2 negative controls, 6 confirmed marijuana exposures, and 13 suspected exposures (Suppl. Table 1). No one breed was overrepresented. Cases were presented to KSU-VHC (5), VECC (10), Animal ER (2), and UI-VTH (2; Suppl. Table 1). The concentration of THC in serum, although different numerically, was not statistically significant (p = 0.25) by institution (data not shown).

In serum, THC was detected in all 19 samples from confirmed and suspected marijuana cases (Table 3). The 2 dogs without exposure to marijuana had undetectable concentrations of the cannabinoids in sera. The median concentration of THC detected was 65.0 ng/mL (0.14–160 ng/mL ± SD 47.6). 11-OH-THC was detected in 16 of 19 serum samples, and the median concentration was 4.8 ng/mL (1.15–17.8 ng/mL ± 5.0). THC-COOH was detected in 17 of 19 serum samples, and the median concentration was 2.18 ng/mL (0.71–7.79 ng/mL ± 2.61). CBD was detected in 12 of 19 serum samples, and the median concentration was 0.28 ng/mL (0.11–82.5 ng/mL ± 23.9). Two samples had >100-fold higher concentrations of CBD than the other 17 samples that were similar in concentration and were 0.1–0.4 ng/mL (Suppl. Table 2). THC-glu was detected in 16 of 19 serum samples, and the median concentration was 2.05 ng/mL (0.72–18.3 ng/mL ± 4.58).

Table 3.

Descriptive statistics for cannabinoid metabolites (ng/mL) detected in canine serum samples.

Metabolite	No. of samples	Mean	Median	SD	Minimum	Maximum
THC	19	68.0	65.0	47.6	0.14	160
11-OH-THC	16	6.12	4.78	5.02	1.15	17.9
THC-COOH	17	3.52	2.18	2.61	0.71	7.79
CBD	12	8.70	0.28	23.9	0.11	82.5
THC-glu	16	4.14	2.05	4.58	0.72	18.3

In urine, no metabolite was detected consistently in all samples (Table 4). THC was detected in 8 of 19 urine samples, and the median concentration was 0.24 ng/mL (0.17–2.74 ng/mL ± 0.94). 11-OH-THC was not detected in any sample. THC-COOH was detected in 16 of 19 urine samples, and the median concentration was 1.84 ng/mL (0.33–16.8 ng/mL ± 5.48). Both CBD and THC-COOH-glu were detected in 2 of 19 urine samples at <1 ng/mL. THC-glu was detected in 15 of 19 urine samples, and the median concentration was 1.23 ng/mL (0.42–11.1 ng/mL ± 3.74). Urine metabolites were also standardized to urine creatinine (Table 5; Suppl. Table 3).

Table 4.

Descriptive statistics for cannabinoid metabolites (ng/mL) detected in canine urine samples.

Metabolite	No. of samples	Mean	Median	SD	Minimum	Maximum
THC	8	0.74	0.24	0.94	0.17	2.74
11-OH-THC	0	ND	ND	ND	ND	ND
THC-COOH	16	4.68	1.84	5.48	0.33	16.8
CBD	2	0.36	0.36	0.27	0.16	0.55
THC-COOH-glu	2	0.29	0.29	0.01	0.28	0.29
THC-glu	15	3.53	1.23	3.74	0.42	11.1

ND = not detected.

Table 5.

Descriptive statistics for cannabinoid metabolite:creatinine ratios detected in canine urine samples (ng metabolite:mg creatinine).

Metabolite	No. of samples	Mean	Median	SD	Minimum	Maximum
THC:Cre	8	0.28	0.22	0.22	0.05	0.74
11-OH-THC:Cre	0	ND	ND	ND	ND	ND
THC-COOH:Cre	16	3.23	1.32	3.76	0.16	11.2
CBD:Cre	3	0.19	0.19	0.10	0.12	0.26
THC-COOH-glu:Cre	2	0.08	0.08	0.05	0.04	0.11
THC-glu:Cre	15	2.75	0.98	3.21	0.25	10.7

ND = not detected.

Of the 21 urine samples tested with the OTC human urine drug test, 20 were negative (Suppl. Figs. 2–21). One urine sample was unable to be run on the OTC human drug test because of an insufficient volume of urine; however, on HPLC-MS/MS, THC-COOH was detected at 2.19 ng/mL, which is well below the 50 ng/mL threshold of the drug screen (Suppl. Table 4). The 2 dogs without exposure to marijuana had undetectable concentrations of the cannabinoids in urine on the OTC test.

All 19 clinical serum samples contained quantifiable concentrations of THC (Table 3), with one sample containing <1 ng/mL THC (above the LLOQ, 0.1 ng/mL; Suppl. Table 2). The primary human metabolite in serum, 11-OH-THC, was consistently detected at >1 ng/mL in all clinical serum samples. In one report, researchers orally administered a cannabis oil extract containing a mixture of 20% THC and 0.5% CBD (1.5 and 0.04 mg/kg, respectively) to 6 dogs and were unable to detect either 11-OH-THC or THC-COOH in serum using HPLC.¹⁰ Other metabolites THC-COOH, CBD, and THC-glu were detected frequently (>50% of samples), but at lower median concentrations. THC-COOH-glu was not included in the serum metabolite profile because of poor recovery, but this was not a problem when quantifying this metabolite in urine. Measuring THC directly in serum of affected dogs revealed an ~10-fold higher median concentration compared to other metabolites in the panel (Table 3). Given that higher concentrations of THC were detected in serum in this group of dogs, utilizing a single serum-based assay with THC as the target analyte could prove more reliable compared to utilizing any other analyte of interest or urine-based assays. However, using a panel of analytes in serum and urine such as in the UPLC-MS/MS method described here will increase the sensitivity of detection because of the variability in canine THC metabolism and time of presentation from time of exposure. The variability in canine THC metabolism is not well understood, but in general it has a long-reported half-life of 30 h.^5,9 In a clinical setting, it may be helpful to interpret the marijuana metabolite panel as a whole in order to arrive at a diagnosis.

Dogs included in our study produced variable and lower concentrations of cannabinoids in urine compared to sera. THC-COOH and THC-glu were most frequently detected in urine (16 of 19 and 15 of 19, respectively; Tables 3, 4). Three dogs with detectable cannabinoids in sera did not have detectable concentrations of cannabinoids in urine. THC was detected in 8 of 19 urine samples at low but quantifiable concentrations. A small-scale, canine marijuana pharmacokinetic study produced in the 1970s reported that only a small amount of THC was eliminated in the urine unchanged, and as a result of enterohepatic recirculation, more complex, conjugated metabolites were eliminated in urine.⁶ Unlike humans, dogs with marijuana exposure do not seem to produce large quantities of THC-COOH in urine. Most OTC human urine drug screens can qualitatively detect THC-COOH when the concentration is >50 ng/mL, which was not sensitive enough for this sample of dogs exposed to marijuana.

Limitations of our study include a small sample size from multiple veterinary hospitals and the inability to accurately record the magnitude of marijuana exposure, type of cannabis product, duration of ingestion to time of presentation, and the reluctance or inability of owners to describe or quantify the ingestion. Our results indicate that a UPLC-MS/MS method can be a useful aid in diagnosing marijuana toxicosis in dogs, whereas using an OTC human urine drug test is not a useful test for confirming marijuana exposure in dogs because of the low content of THC-COOH in urine.