Quantification of Δ9-tetrahydrocannabinol in urine as a marker of cannabis abuse

Cannabis is illegal, but still, its prevalence is remarkable across the social and religious ritual landscape of India. Cannabis’ medical and religious implications have a long history of cultivation and consumption that has now reached its peak efflorescence1. According to the United Nations Office on Drugs and Crime2, New Delhi and Mumbai in India are two of the top ten cities in the world that consume the most cannabis3.

(−)-trans-9-tetrahydrocannabinol (THC) is the most abundant phytocannabinoid responsible for the majority of the psychoactive effects. THC concentration varies depending on which part of the plant was used to make the stuff, and it can be used to classify cannabis. Dried leaves (Bhang) have the least amount of THC (1-2% by weight). Ganja/weed/marijuana is a mixture of dried flowers and leaves adjacent to them that contains 10-12 per cent of THC. The remaining cannabis products are hashish (20-30% THC) and hash oil (40-60% THC)4. Cannabis products are addicting and can lead to psychological and medical problems5. A recent nationwide study has reported that 2.8 per cent of adults in India use cannabis products, out of which 1.2 per cent use illicit preparations Alarmingly, one in four cannabis users67 exhibits problematic use. Even though all cannabis-related products, except bhang/leaves, are still banned under the Narcotic Drugs and Psychotropic Substances (NDPS) Schedule I, there are approximately 7.2 million cannabis users in India, second only to alcohol users which number 57 million people7.

Biochemical detection of cannabis use is important in the medical and forensic contexts8. Since THC is the primary psychoactive compound in cannabis products, biochemical testing aims to identify evidence of its presence in the human body. However, THC is rapidly metabolized by the liver, and therefore, biochemical tests are designed to detect the presence of THC metabolites in various specimens such as urine, hair, blood and sweat. 11-nor-Δ9-tetrahydrocannabinol-carboxylic-acid (THC-COOH) is the end metabolite of THC oxidation and is well suited for biochemical detection and quantification5.

Methods for biochemical detection of cannabis use can be broadly classified into qualitative (screening) and quantitative (confirmatory) tests. Qualitative tests give a positive/negative result and are used as a point of contact test. Due to high false-positive rates, these tests should be supplemented with quantitative tests to confirm the presence of THC-COOH9. Some clinical concerns cannot be resolved only by qualitative tests, even if we ignore the problem of false-positive results. A positive qualitative test, for instance, can reveal single usage within the previous three days or years of consistent heavy daily use10. If the healthcare staff becomes aware of a patient’s obvious intoxication and inebriation on the seventh day and the screening is positive, the doctor will not be able to identify whether it indicates recent intake or carryover. Confirmatory tests are essential to make these differences and measuring the concentration of THC-COOH in urine becomes essential11.

Quantification of THC-COOH in urine samples presents several challenges. THC-COOH is excreted as a glucuronide, and therefore, enzymatic or alkali-induced lysis is required before analysis. The second challenge is handling samples with organic solvents such as hexane (central nervous system depressant). Team members involved in THC-COOH, should not be regularly exposed to such solvents and prepare samples in fume hoods. The third challenge is derivatization that makes the analyte suitable for gas chromatography and mass spectrometry (GC-MS). The last challenge is to ensure that the analyte is optimally concentrated, interference-producing chemicals in the matrix are minimized and the sample can tolerate the high temperature of the GC injection port8. Although technological advancement is beneficial, efforts should also take into account the limits of routine clinical laboratories. An ideal sample preparation protocol should be fast, time and cost-efficient, suitable for batch processing, use non-toxic chemicals and have minimum sample transferring steps. Even more importantly, when combined with GC-MS, the protocol must give valid and replicable results. In this work, we present a method that meets many of these requirements and has been validated in the Indian context.

Material & Methods

This study was carried out at the toxicology services attached to the Centre for Addiction Medicine (CAM), National Institute of Mental Health and Neurosciences, Bangaluru, Karnataka, from October 2018 to March 2019. The study was approved by the Institutional Ethics Committee.

CAM provides outpatient, inpatient and emergency services to patients with substance use disorders; urine drug testing is carried out as protocolized for specific clinical scenarios. In brief, clinicians request urine drug testing through a requisition slip that lists the self reported substance use, date of last use and sociodemographic details. All samples undergo a qualitative test for the common groups of substances at the laboratory, followed by a quantitative test. The sampling frame for this study consisted of all toxicological tests conducted during the study period (except tests done only to detect alcohol biomarkers) and investigation involved de-identified, aggregated laboratory test data.

Patients were instructed to collect 5-10 ml of urine in pre-labelled 20 ml leak proof sterile plastic containers. These samples are immediately transferred to the laboratory and cross-matched with the requisition details entered by clinicians. Sample adulteration was assessed using the commercial adulteration strips (ABON and INC), which assess samples based on specific gravity, pH, creatinine and nitrite concentration and presence of glutaraldehyde/oxidant/pyridinium chlorochromate12. In most cases, the samples obtained were processed on the same day. In case of delay, the samples were stored in a refrigerator at 4°C for not more than 48 h. Initial screening was done for multidrug use (cannabis, morphine, benzodiazepines, cocaine and amphetamine) by commercial test kits (Abon Inc., NY, USA) according to the manufacturer’s specifications13. Samples that tested positive for THC-COOH were then analyzed further.

Peak integration was performed in Agilent Chemstation version (A.01.03, Santa Clare, CA, USA) software. GraphPad Prism 7 (GraphPad Software, San Diego, CA, USA) was used for regression and statistical analysis, respectively and all illustrations were done with the same software as well.

Standards and reagents: Standards of 11-nor-9-carboxy-∆-9THC (Carboxy-THC) and internal standards 11-nor-9-carboxy-∆-9THC D-3 (THC-COOH D₃) were procured from Cerilliant (Sigma Aldrich, USA). Organic solvents acetonitrile and methanol ultra-gradient (LCMS grade) were supplied by Sigma Aldrich. Ultrapure water was obtained from, Ultra-Pure Water System (Millipore, USA). All working standards of 100 µg/ml strength were prepared in methanol and stored at −20°C. For centrifugation, Spin-win 02 microcentrifuge was used. Control urine lyophilized samples were purchased from Medi-chem Germany.

Preparation of calibration standards: Stock solutions containing Carboxy-THC and internal standards THC-COOH D₃ were prepared by diluting the methanolic reference solutions to obtain the intermediate solution (10,000 ng/ml) and the working solution (1000 ng/ml). These solutions were stored at 2-8°C in amber bottles until use. Calibration standards were prepared in drug-free urine at concentrations of 15, 30, 75, 150 and 300 ng/ml. The established calibration curve ranging from five to 300 ng/ml (with 2 ml of urine sample) was verified by running the three levels of commercial quality control materials (Biorad Liquichek™ Urine Toxicology Control) and six levels of THC-COOH linearity samples run in triplicate.

Urine sample preparation: To urine samples (2 ml) 50 µl of internal standard (1 µg/ml) and 500 µl of 6M NaOH were added to adjust at pH 13±.5; after vortexing (1-2 min) sample was placed in water bath at 80°C for 30 min to free THC-COOH from glucuronide finally pH was adjusted to 4.0±0.50 with glacial acetic acid before Solid Phase Extraction (SPE). Bond Elute cartridges were conditioned with methanol (1 ml), before adding sample, followed by washing methanol and distilled water (1:1). One mililitre of hexane and ethyl acetate (1:3) was used for elution and drying was done using the vacuum evaporator (Gevac EZ-2, Marshall Scientific, Hampton, NH, USA).

Derivatization: Silylation was used to prevent decarboxylation of cannabinoid acids when exposed to high temperatures of GC-MS. Vacuum dried samples were reconstituted with 100 µl of Bis N-methyl -N-tri-methyl-silyl-trifluoroacetamide with one per cent trimethylchlorosilane (BSTFA) and incubated for half an hour at 80°C. Samples were transferred to vials at room temperature and injected to GC-MS, cannabis metabolites were determined by comparing the responses of the unknown sample to the responses of the calibrators.

Instrumentation: The analysis was completed on Agilent 7890A gas chromatograph and 5975C mass selective detectors (GC-MS) equipped with 7693 auto-sampler (©Agilent Technologies, Inc., Wilmington, DE, USA). Chromatographic separation was achieved on nonpolar and low bleed column from Agilent (122-5532). The injector and transfer line temperature were between 250 and 280°C and the mass selective detector operated in the electron impact mode at 70 eV and ion source temperature of 150°C. Injection port temperature was 200°C and helium was carrier gas. Initial oven temperature was 100°C and for two minutes. Followed by an increase at the rate of 35°C/min to 200°C, then 23°C/min till temperature reached to 250°C, and finally, increase was made at the rate of 20°C/min till 300°C. Data were acquired using Chemstation software (©Agilent Technologies, Inc.).

The trimethylsilyl derivatives were identified with masses THC-COOH m/z: 473 was target ion and 371 and 488 were qualifiers (Figs 1 and 2). For internal standards, THC-COOH m/z 476.1 target ion and 374.2 were qualifiers. For IS THC-COOH D₃ qualifier was 374,476 and RT was 11.7 min. Molecular justification of these ions is presented in Figure 3.

Close

Fig. 1

Mass spectrum showing m/z of THC-COOH. THC-COOH, 11-nor-Δ9-tetrahydrocannabinol-carboxylic-acid

Export to PPT

Close

Fig. 2

Graph showing a sequential decline in THC-COOH levels among patients.

Export to PPT

Close

Fig. 3

Chromatogram showing all three masses of THC-COOH.

Export to PPT

The following characteristics were used for method validation14 :

• (i) Calibration model: The ratio of area under the curve of five THC-COOH standards and the internal standard was analyzed with a least squares regression model to arrive at a calibration model

• (ii) Precision: Intra- and inter-day (three consecutive days) precision was calculated at low (10 ng/ml), medium (100 ng/ml) and high (250 ng/ml) concentrations. Recovery was also calculated at these levels

• (iii) Linearity: To assess linearity, standards were prepared by spiking urine samples with increasing concentration of THC-COOH. The assay was deemed linear until one or more of the qualifying ion ratios failed or the determined and expected differed by more than 10 per cent

• (iv) Limits of quantification and detection: The limit of quantification (LOQ) and limit of detection (LOD) were defined as the lowest analyte (THC-COOH) concentration that can be detected against signal-to-noise ratio 10:1 and 3:1 in spiked urine triplicate.

• (v) Matrix interference and carryover effects: Six blank matrices from different subjects were used to check matrix interference. Blank matrix samples were analyzed after high concentration spiked samples were analyzed to quantify carryover effects.

• (vi) Method robustness: small changes in sample preparation such as derivatization time, sample volume and pH were introduced to assess robustness. Two different operators analyzed the same samples, and root square deviation was calculated to assess robustness to operator characteristics.

Stability: Stability of the target analytes was studied under the specific conditions and time intervals designed to imitate for biological sample collection and storage. Short-term stability, stability after freeze/thaw cycles and stability in processed samples were evaluated in triplicate at QC levels (10, 100 and 250 ng/ml) and coefficients of variation (CVs) were calculated.

Results

The assay was linear in the calibration range of 15-300 ng/ml (r² of 0.99) with two ml of sample; LOD was 5.0 ng/ml and LOQ was 7.0 ng/ml. For three days in a row, inter-day accuracy was evaluated. Intra- and inter-day precision was 8.69 and 10.78 per cent, respectively (Table) and per cent recovery for the lowest concentration was 97.5. Matrix effect was investigated using six lots of blank matrix form individual donors, which was ≤10 per cent while test carryover was negligible. In the robustness analysis, the accuracy of 89.9 per cent was well within the range (80-120%). The bias and coefficient of variation were <15 per cent. During the stability check of the processed samples after 24 h (on auto-sampler), CV ranged from 0.15 to 12.22 per cent. Short-term stability of spiked samples extracted at various time intervals (0, 24, 48 and 72 h) and compared to freshly prepared samples, CVs obtained were ≤10 per cent. After three freeze/thaw cycles, the CVs obtained were ≤10 per cent. The current analytical approach was linear with a calibrator’s accuracy [mean relative error (bias) between the measured and spiked values] within a ±15 per cent range for all concentrations. CVs typically lower than 15 per cent were obtained for precision.

A total of 939 urine samples relevant to this study were analyzed in the laboratory during the study period. Of these 939 samples, 213 tested positive for THC-COOH presence in the screening test and constituted the final sample for this study. Out of these 213 samples, 128 (60%) were positive only for cannabinoids, 70 (33%) for cannabinoids and sedatives, 14 (6.5%) for cannabinoids and morphine and one tested positive for cannabis, morphine and cocaine. These 213 samples came from 195 adult male individuals with mean (± standard deviation) age of 40.6±14.2 yr.

The current work also contained data of twelve participants representing a successive decrease in THC-COOH levels; among them THC-COOH levels ranged between 50 and 1422 ng/ml on the first analysis (Fig. 3). The second analysis was performed after 9-15 days depending on the follow up, which showed a decline in THC-COOH levels. This analysis had values 29-50 ng/ml for three subjects, whereas the remaining had values from 51 to 1172 ng/ml (Fig. 3). During data analysis, one participant presented with a considerable increase in THC-COOH values compared against the results obtained following the first analysis (56 ng/ml to 1172 ng/ml). Examining the case history of this participant indicated a relapse. The remaining samples reported a gradual decrease in the THC-COOH levels in the consecutive follow ups after 7-15 days. In the third analysis, THC-COOH values reached below the cutoff (12-44 ng/ml) within a month or so (30-40 days). The findings demonstrated that the current analytical tool was able to effectively evaluate and track the patterns of abstinence and relapse.

Discussion

Quantifying self reported use of alcohol and nicotine are simple, with people stating the number of alcoholic beverages they drank and the number of cigarettes smoked15. In contrast, cannabis abuse is more difficult to detect and quantify since it may be smoked in a variety of ways, including4 joints, blunts, pipes, bongs and vaporizers, all of which can be shared. A significant variance in 9-THC potency (2-20%) and availability of very potent forms (schwag and sinsemilla) further complicate this issue16. There are numerous analytical methods for analyzing THC-COOH in biological fluids. The method chosen is determined by the infrastructure available and the anticipated amount of samples to be analyzed.

To determine the total THC-COOH in the urine sample, the glucuronide must first be hydrolyzed, either enzymatically or by alkaline hydrolysis. However, the majority of the enzyme hydrolysis methods outlined are time consuming (8-16 h)17. Alkaline hydrolysis with sodium hydroxide is not only time efficient and repeatable but also cost-effective. SPE was performed on pretreated samples using C18 disc extraction cartridges, providing a clean extract with no interference. This approach is suitable for use in laboratories that process a high number of urine samples for THC metabolite analysis by GC-MS due to its simplicity and repeatability18. In previous studies, methyl iodide and tetra-methyl ammonium hydroxide were used to derivatize dimethyl sulfoxide (DMSO)1920. Dermal exposure to DMSO has been linked to pruritis, urticaria, erythema, exfoliation and pigmentation, while iodomethane is a known nephrotoxin and neurotoxic as well as a potential carcinogen. The most common derivatization process is silylation with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA), which quickly volatilizes the sample and is thus ideal for non-volatile materials for GC analysis. N-methyl-N-(trimethylsilyl)trifluoroacetamide21 and BSTFA22, when combined with trimethylchlorosilane or ethyl acetate2324, improve the sensitivity and resolution of cannabinoids in extract.

Chromatographic separation was achieved on DB-5 MS column having (5%-Phenyl)-methylpolysiloxane coating. This is non-polar, low bleed, bonded, crosslinked, and has a high temperature tolerance which provided accurate identification and quantification, with excellent selectivity and inertness25. Helium proved to be a faster and safer carrier gas than nitrogen, and it has a higher resolution and lower background noise than hydrogen. The running cost would be around ₹ 1500 for any new facility. Anderyak et al25 evaluated previously screened urine samples that were positive for cannabis and found that the average recovery rate was around 95 per cent and that the current method had a good per cent recovery of 97.5 per cent with low concentration. The SPE and GC-MS combination is appropriate and efficient for batch processing results in less than a day. The use of an SPE resulted in a clean extract with minimum chemical exposure and without interference from other substances. THC-COOH quantification can also be used to track cannabis use and may help in management.

This study however, had a limitation as the current approach is incapable of distinguishing between acute and chronic cannabis use. Future studies could look at the active metabolite THC to gain a better picture of recent use.