Does acute stress induced via cold water immersion increase blood THC concentrations in regular cannabis users?

Danielle McCartney; Jordan Levoux; Rebecca Gordon; Laura Sharman; Katie Walker; Jonathon C Arnold; Iain S McGregor

doi:10.1007/s00213-025-06833-8

. 2025 Jun 14;242(12):2785–2799. doi: 10.1007/s00213-025-06833-8

Does acute stress induced via cold water immersion increase blood THC concentrations in regular cannabis users?

Danielle McCartney ^1,^2,^3,^✉, Jordan Levoux ^1,², Rebecca Gordon ^1,^2,⁴, Laura Sharman ^1,^2,⁴, Katie Walker ^1,^2,⁴, Jonathon C Arnold ^1,^2,⁴, Iain S McGregor ^1,^2,³

PMCID: PMC12675739 PMID: 40515781

Abstract

Rationale

Δ⁹-tetrahydrocannabinol (THC), the principal intoxicant in cannabis, is a highly lipophilic substance that is readily stored in fat. Preclinical research suggests that fat-stored THC can be released into circulation under lipolytic conditions, such as stress. However, compelling evidence of this phenomenon is lacking in humans.

Objectives

Here, we investigated the effects of acute stress on blood (plasma) THC (and 11-COOH-THC) concentrations as well as indices of intoxication (e.g., cognitive function, subjective drug effects) in ‘regular’ (i.e., > 3 days/week) cannabis users.

Methods

Fifteen volunteers (n = 9 female; cannabis use: 5.0 ± 1.5 days/week) participated in a single-arm (i.e., pre-/post-intervention) trial. The intervention (stressor) was cold water immersion (CWI); specifically, 10 min at ~ 10 °C. Plasma cannabinoid concentrations, cognitive function and subjective drug effects were measured pre-, shortly post- and 2 h post-intervention. Measures of stress (e.g., heart rate [HR], blood pressure [BP], subjective ratings) and lipolysis (i.e., plasma free fatty acid [FFA] and glycerol concentrations) were also obtained.

Results

CWI produced a small but significant stress response, characterised by increased HR and systolic BP and decreased subjective “calmness”. It also increased plasma FFA and glycerol concentrations, albeit modestly. However, neither plasma THC nor 11-COOH-THC concentrations increased from pre- to post-intervention. Cognitive function also remained unchanged, while subjective drug effects were negligible.

Conclusions

Stress induced via brief CWI does not appear to increase blood cannabinoid concentrations or induce intoxication in moderate cannabis users.

Keywords: Cannabinoid, Plasma, Oral fluid, Biomarkers, Drug testing, Cognition, Psychomotor, Impairment, Ice bath

Introduction

Over 200 million people use cannabis each year (United Nations Office on Drugs and Crime 2023). Some, to treat medical conditions such as chronic pain, insomnia, and anxiety– others, to induce euphoria and relaxation (Lintzeris et al. 2022; MacPhail et al. 2022). Indeed, cannabis, specifically its principal phytocannabinoid, Δ⁹-tetrahydrocannabinol (THC), is a well-known intoxicant that causes dose-dependent subjective and physiological changes, including cognitive and psychomotor impairment, in users (McCartney et al. 2021; Wickens et al. 2022).

THC is a highly lipophilic substance with poor aqueous solubility (Banister et al. 2019; Thomas et al. 1990). This means that while most THC (e.g., ~ 80–90%) is excreted within five days of use as hydroxylated and carboxylated metabolites (e.g., 11-OH-THC, 11-COOH-THC), significant quantities are taken up and stored in fat deposits throughout the body (Chayasirisobhon 2020). Indeed, THC has been detected in human fat biopsies 28 days following cannabis use (Johansson et al. 1989). Of course, fat is a biologically active tissue that releases lipids (e.g., free fatty acids [FFAs]) and glycerol into the blood, through a process known as lipolysis, whenever the body requires energy (Edwards and Mohiuddin 2020).

Given that THC is stored in fat, and lipolysis can be triggered by stress; specifically, the release of stress hormones such as adrenocorticotrophic hormone (ACTH), adrenaline and noradrenaline (Edwards and Mohiuddin 2020), scientists have theorised that physical and psychological stressors might increase blood THC concentrations in regular cannabis users (Gunasekaran et al. 2009; Westin et al. 2014; Wong et al. 2013, 2014). This is significant as blood THC concentrations are often used as evidence in medicolegal cases involving highly stressful and/or traumatic events (e.g., motor vehicle accidents, assaults).

This ‘reintoxication hypothesis’ has some scientific support. First, there have been two published cases of anomalously high post-mortem blood THC concentrations (i.e., 90 and 290 ng/mL) in victims of traumatic death (Collins et al. 1997). Second, the administration of ACTH has been shown to increase: (a) the THC concentration of a medium containing rat adipocytes (pre-treated with THC); and (b) blood THC concentrations in rats (pre-treated with THC) (Gunasekaran et al. 2009). Third, stressors such as exercise and food deprivation have been shown to increase blood THC (Gunasekaran et al. 2009) and 11-COOH-THC (Gunasekaran et al. 2009; Wong et al. 2014) concentrations in rats (pre-treated with THC). To date, however, only two studies have investigated the phenomenon in humans.

Wong et al. (2013) measured the plasma THC and 11-COOH-THC concentrations of 14 ‘regular’ (i.e., ≥ 5 days per week) cannabis users before, after, and 2 h after 35 min of cycling at 50% of aerobic capacity (VO_2max) in either a fed (n = 7) or fasted (n = 7) state. The study showed that exercise (but not fasting) had a strong (Cohen’s d_z = 1.16) and significant (transient) effect to increase plasma THC concentrations. A positive correlation between the exercise-induced change and body mass index (BMI) was also observed, suggesting that individuals with greater fat mass might be more vulnerable to the effect.

Meanwhile, Westin et al. (2014) measured the serum THC and 11-COOH-THC concentrations of six daily cannabis users before and after: (a) 45 min running at 60–75% of age-predicted heart rate (HR) maximum; and (b) 24 h of food deprivation. No “major differences” were said to be observed. However, a considerable proportion (≥ 80%) of participants still had mildly elevated serum THC and 11-COOH-THC concentrations after exercise and fasting, respectively. Further research clarifying whether acute stress increases blood cannabinoid concentrations in regular cannabis users is, therefore, warranted.

The primary aim of this study was to investigate the effects of acute stress on blood (plasma) THC and 11-COOH-THC concentrations in regular cannabis users. The specific stressor employed was cold water immersion (CWI), as cold exposure appears to trigger the release of stress hormones (Eimonte et al. 2022) and induce lipolysis (Straat et al. 2022). The secondary aim was to determine whether a stress-induced change in plasma THC concentration was sufficient to: (a) induce mild intoxication, as indexed by subjective reports and objective measures of cognitive function, and (b) elicit positive point-of-care (POC) (e.g., roadside, workplace) oral fluid drug tests. We hypothesised that CWI would increase plasma THC and 11-COOH-THC concentrations, inducing mild intoxication. The likelihood of a stress-induced increase in plasma THC concentration affecting oral fluid THC concentrations was expected to be low.

Methods

Study design

A single-arm (i.e., pre-/post-intervention) non-clinical trial was conducted at the University of Sydney. The trial was approved by the University’s Human Research Ethics Committee (2023/269) and conducted in accordance with the Declaration of Helsinki (1983).

Participants

Eligible volunteers were: (1) aged between 18 and 40 years; (2) ‘regular’ (i.e., ≥ 3 days/week for ≥ 3 months) users of cannabis or THC-containing products (as evidenced by a positive POC urine drug test for 11-COOH-THC); and (3) proficient in English.

The following exclusion criteria applied: (1) hypersensitivity to cold (e.g., Raynaud’s syndrome); (2) recent (i.e., ≤ 3 months prior) involvement in a ‘temperature manipulation’ (i.e., cold stressor) program; (3) BMI < 22.0 kg/m²; (4) cardiac or vascular disease (including hypertension); (5) haemophilia; (6) high resting systolic blood pressure (SBP) (i.e., > 140 mmHg); (7) asthma; (8) a neurological disorder or intellectual disability; (9) open wounds; (10) unwilling or unable to adhere to the study procedures (including the standardisation procedures outlined in Sect. 2.5.1); and (11) pregnant or lactating.

Participant enrolment

Each volunteer completed a short online questionnaire and a telephone interview. Those suitable to continue then scheduled a face-to-face screening. Here, they were informed about the nature and risks of the experimental procedures before providing written informed consent and being assessed for eligibility via self-report, having their height, weight and blood pressure measured, and completing a POC urine drug test (NxStep OnSite Drug Test, DrugCheck^®, MN, USA). A positive test (i.e., suggesting that urinary 11-COOH-THC concentrations were > 50 ng/mL) was taken as evidence of regular cannabis use. Eligible participants then provided basic demographic information, had their fat mass measured via standard procedures (BC-730 Body Composition Monitor, Tanta^®, WA, Australia), and practised the cognitive and psychomotor tasks they would complete at their test session.

Intervention

The intervention (stressor) was CWI. Participants sat in a portable bath (sub-merged to the level of their clavicle) in light clothing or swimwear for a period of 10 min. The bath contained ~ 165 L of cold tap water and ~ 20 kg of ice. The temperature (~ 10 °C) was measured at 2.5-minute intervals throughout CWI (THERM-cert, Brannan^®, England). This protocol has been shown to trigger the release of hormones (i.e., adrenaline, noradrenaline, cortisol) (Eimonte et al. 2022) that can be expected to induce lipolysis (Straat et al. 2022).

Test sessions

Each participant completed one test session at the Brain and Mind Centre (BMC).

Standardisation procedures

Participants were instructed to: (1) abstain from cannabis and caffeine (> 12 h); (2) abstain from alcohol and other psychoactive substances (> 24 h); (3) spend > 8 h in bed overnight; (4) fast overnight (i.e., > 8 h) (as fasting has been shown to potentiate stress-induced increases in plasma FFA concentrations (Iwayam et al. 2020)); and (5) consume ~ 500 mL of water (i.e., on waking) prior to their test session.

Experimental procedures

Participants arrived at the laboratory (in a fasted state) between ~ 07:00–11:00 AM and were asked to indicate whether they complied with each of the standardisation procedures. Compliant participants then provided information about their last cannabis use, re-familiarised themselves with the cognitive and psychomotor tasks, and completed a series of pre-intervention assessments (outlined in Sect. 2.6 ‘Data Collection’) before undergoing CWI. The pre-intervention assessments were repeated shortly post-intervention (i.e., after participants had towel-dried and dressed themselves) and 2 h post-intervention. Individuals were also asked to indicate whether they experienced any “unfavourable signs or symptoms” (adverse events) before leaving the laboratory. Participants remained fasted (i.e., were only permitted to consume plain water) throughout the (~ 3-hour) test session.

Data collection

Primary outcome

Plasma cannabinoid concentrations were measured pre-, ~ 5 min post- and 2 h post-intervention using ultra-high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS) and previously validated methods (Kevin et al. 2021). Samples were analysed in triplicate, with those derived from the same participant analysed on the same plate.

Blood was collected into pre-treated EDTA vacutainers (Beckton Dickson, NJ, USA). The vacutainers were centrifuged at 2500 g and 4 °C for 15 min within 30 min of collection and the plasma supernatant was stored at − 80 °C until analysis.

Secondary outcomes

The following assessments were conducted pre-, post- and ~ 2 h post-intervention:

Cognitive and psychomotor function were measured using three computerised tasks: (1) the Digit Symbol Substitution Task (DSST) (1.5 min); (2) the Divided Attention Task (DAT) (~ 4 min); and (3) the Paced Serial Addition Task (PSAT) (~ 4 min). These tasks have previously demonstrated sensitivity to the impairing effects of THC and are detailed elsewhere (Arkell et al. 2019b, 2020).
Subjective intoxication (i.e., “stoned”, “strength of drug effect”, “euphoric”) was measured using 100 mm visual analog scales (VASs), where zero = “not at all” and 100 = “extremely”.
A POC oral fluid drug test; specifically, the Dräger DrugTest^® 5000 (DT5000), was administered as per the manufacturer’s instructions (Arkell et al. 2019a; McCartney et al. 2022; Suraev et al. 2024). The DT5000 generates a qualitative result (i.e., positive, negative, invalid) for cannabis, cocaine, opiates, benzodiazepines, amphetamines and methamphetamine, and has an adjustable limit of detection (LOD) (i.e., for THC) that was set to 25 ng/mL for the purpose of this experiment.
Oral fluid THC concentrations were measured using LC-MS/MS and previously published methods (Arkell et al. 2019a; McCartney et al. 2022; Suraev et al. 2024). Samples were analysed in triplicate, with those derived from the same participant analysed on the same plate. Oral fluid was collected using the Quantisal^™ device (Immunalysis Corporation, CA, USA) and stored at + 4 °C (in the device) until analysis (within 90 days of collection (Scheidweiler et al. 2017).

Measures of stress

The following were measured pre-, post- and ~ 2 h post-intervention:

SBP, diastolic blood pressure (DBP) and heart rate (HR) (in duplicate) using an automated sphygmomanometer (HEM-7121 J, Omron, Kyoto, Japan).
Subjective stress (i.e., “calm”, “bored”, “nervous”, “excited”) using 100 mm VASs, where zero = “not at all” and 100 = “extremely”.
Tympanic temperature using a thermometer (DET-1015, Welcare, Hangzhou, China).

The following were measured at 2.5-minute intervals throughout CWI:

Tympanic temperature (as above).
HR using a wristwatch (Vivoactive 4 S, Garmin, KN, USA).
Affective valence using the 11-point Feelings Scale, where − 5 = “feeling very bad” and + 5 = “feeling very good” (Hardy and Rejeski 1989).

Measures of lipolysis

Plasma FFA and glycerol concentrations were measured pre-, ~ 5 min post- and 2 h post-intervention using commercial kits (LabAssay NEFA Kit (Cat No. 633–52001), Wako Chemical, Osaka, Japan; Glycerol Colorimetric Assay Kit (Cat No. 10010755), Cayman Chemical, MI, USA). Samples were analysed in triplicate, with those derived from the same participant analysed on the same plate. (Blood was collected, processed, and stored as described in Sect. 2.6.1 ‘Primary Outcome’).

Other descriptive measures

Urinary cannabinoid concentrations were measured (in triplicate) using LC-MS/MS and previously published methods (Kevin et al. 2017). A pre-intervention urine sample was collected into a sterile plastic cup and stored at − 80 °C until analysis.

Statistical analysis

Continuous variables were analysed using linear mixed effects models that included Time (within-subject categorical: pre-, post-, and 2 h post-intervention) as a fixed effect, with Participant as a random effect. Sex (categorical: male and female), Fat Mass (continuous), the Time by Sex interaction, or the Time by Fat Mass interaction was also included as a fixed effect if it reduced the Akaike information criterion (AIC) value of the model. If the residuals were non-normally distributed (Shapiro-Wilk test, p < 0.05) or heteroscedastic (Breusch-Pagan test, p < 0.05), the dependent variable was square-root transformed and re-analysed (and if unimproved, log transformed). If the variable included zeros, a small positive constant (half the smallest non-zero value) was added prior to transformation. If neither transformation was curative, a gamma generalised linear mixed-effects model (with a log link) was substituted.

Count variables (e.g., certain measures of cognitive function) were also analysed using generalised linear mixed-effects models. These models included the same fixed and random effects as those described above but were fitted to a Poisson distribution (as no data were zero-inflated or over-dispersed).

Two-sided (unadjusted) post hoc comparisons were performed to compare the estimated marginal means if a significant main or interaction effect was observed. Statistical significance was accepted as p < 0.05. All values presented in text are estimated marginal means and their 95% confidence intervals (CIs), unless otherwise stated.

All statistical analyses were performed in R version 4.2.0 using the following packages: ‘tidyverse’ (mutate function) (Wickham et al. 2019), ‘car’ (Anova and qqp functions) (Fox and Weisberg 2019), ‘MuMIn’ (AICc function) (Bartoń 2022), ‘lme4’ (lmer function) (R Core Team 2022), ‘lmerTest’ (Kuznetsova et al. 2017), ‘glmmTMB’ (glmmTMB function) (Brooks et al. 2017), ‘emmeans’ (emmeans function) (Lenth 2023) and ‘performance’ (check normality, check heteroscedasticity, check model functions) (Lüdecke et al. 2021); ggplot2 (Wickham 2016) and ‘Rmisc’ (Hope 2022) were also used to generate plots.

Sample size

Wong et al. (2013) found that exercise increased plasma THC concentrations (Cohen’s d_z = 1.16) in regular cannabis users. Using a power (1-β) of 0.8, a two-sided α of 0.05, and a more conservative Cohen’s d_z of 0.80, we predicted that 15 participants would be required to test the primary hypothesis. A target sample size of 18 was selected to account for attrition.

Results

Recruitment and retention

Eighteen participants signed informed consent between August 4 and October 10, 2023 (Fig. 1). All were eligible to participate and attended a test session. However, only 17 were exposed to the intervention. The remaining individual was sent home pre-intervention (and excluded from the final sample) after he reported using psychoactive substances (benzodiazepines and amphetamines) < 24 h prior to arrival. A second individual was also excluded after being found to have high oral fluid THC concentrations (i.e., 33 ng/mL), inconsistent with the required 12 h of cannabis abstinence (Lee et al. 2015; Odell et al. 2015). All other participants had oral fluid THC concentrations ≤ 6 ng/mL. Finally, a third participant who (a) returned a positive (pre-intervention) oral fluid drug test for cocaine but (b) was not sent home (as he affirmed that he had not used cocaine < 24 h prior to arrival), was excluded on the basis that residual oral fluid cocaine concentrations seldom exceed 20 ng/mL (Cone and Weddington 1989) (i.e., the LOD on the DT5000 (Gjerde et al. 2018). The 15 remaining participants were included in the final (analytical) sample.

Participant characteristics

The demographic characteristics of the participant population are summarised in Table 1. Briefly, individuals reported using cannabis an average of 5.0 ± 1.5 days/week (mean ± SD) and ~ 23 [14–40] hours (median [interquartile range, IQR]) prior to their test session. They also had an average BMI and body fat content at the lower end of the ‘overweight’ range (Jeukendrup and Gleeson 2004).

Table 1.

Baseline characteristics (n = 15)

Characteristic	Frequency (n), Mean ± SD or Median [IQR]
General:
Sex (n)
Male	6 (40%)
Female	9 (60%)
Gender (n)
Man	6 (40%)
Women	8 (53%)
Non-binary	1 (7%)
Age (years)	29 ± 4
Height (m)	1.71 ± 0.13
Weight (kg)	75 ± 15
BMI (kg/m²)	25.1 [22.8–26.4]
Body Fat (%)
Males	22.4 ± 5.6
Females	32.9 ± 9.1
All	28.7 ± 9.3
Cannabis Use:
Frequency of use (days/week)	5.0 ± 1.5
Duration of regular use (years)	2.0 [1.0–5.0]
Urinary THC (ng/mL)	1.52 [1.17–1.74]
Urinary 11-OH-THC (ng/mL)	1.15 [0.51–1.90]
Urinary 11-COOH-THC (ng/mL)	26 [10–51]
Perceived potency (n)
Very strong	4 (27%)
Strong	9 (60%)
Average	2 (13%)
Weak	0
Very weak	0
Usual route of administration (n)
Inhalation	12 (80%)
Oral ingestion	2 (13%)
Tinctures	1 (7%)
Reason(s) for use (n)
Recreational purposes	5 (33%)
Medicinal purposes	2 (13%)
Both purposes	8 (53%)

Open in a new tab

Values are frequency (n), mean (95% CI) or median [IQR], as appropriate (i.e., where data are normal and non-normal, respectively). BMI: Body Mass Index; IQR: Interquartile Range; SD: Standard Deviation

Intervention characteristics

Zero, 2.5, 5.0, 7.5 and 10 min into CWI, the water temperature was 11.6 [10.9–11.6], 10.5 [10.1–11.1], 10.7 [10.0–11.2], 11.1 [10.3–11.4] and 11.2 [10.6–11.4]°C, respectively (median [IQR]).

Primary outcomes

Three of the 45 blood samples (2 × post- and 1 × 2 h post-intervention) could not be collected. THC and 11-COOH-THC were present (i.e., above the LOD, 0.1 ng/mL) in all 42 of the remaining samples.

Plasma THC concentration (Table 2; Fig. 2A) did not demonstrate a significant main effect of Time (Table 3). However, plasma 11-COOH-THC concentration (Table 2; Fig. 2B) did (Table 3). Post hoc comparisons showed that concentrations were higher pre-intervention (16 (5, 32) ng/mL, p = 0.002) and post-intervention (17 (6, 34) ng/mL, p < 0.001) than 2 h post-intervention (13 (4, 29) ng/mL) – but did not differ pre- and post-intervention (p = 0.215).

Table 2.

Descriptive statistics

	Baseline	Post-Intervention	2 h Post-Intervention
Primary Outcomes ^{a, b}
Plasma Cannabinoid Concentrations
Plasma THC (ng/mL)	0.70 [0.51–1.62]	0.63 [0.46–1.69]	0.59 [0.35–1.51]
Plasma 11-COOH-THC (ng/mL)	8.2 [4.9–19.4]	7.8 [5.9–24.8]	6.5 [3.9–18.0]
Secondary Outcomes ^c
Digit Symbol Substitution Task
Correct Responses (n)	43.5 [39.5–49.0]	45.5 [39.8–50.8]	45.0 [43.5–48.0]
Attempts (n)	45.5 [42.0–50.8]	46.0 [43.0–50.8]	47.0 [45.5–50.5]
Response Accuracy (%)	97.3 [92.4–99.6]	99.0 [97.7–100]	97.8 [93.9–99.2]
Divided Attention Task
Hits (n)	24.0 [23.0–24.0]	23.0 [22.0–24.0]	23.0 [21.0–23.0]
Response Time (ms)	920 [778–1150]	1033 [911–1188]	1081 [885–1307]
Tracking Error (pixels)	11.7 ± 3.5	11.4 ± 2.5	12.3 ± 3.0
Paced Serial Addition Task
Correct Responses (n)	47.8 ± 8.3	44.9 ± 8.8	44.3 ± 9.6
Response Time (ms)	1367 ± 184	1321 ± 130	1334 ± 135
Subjective Intoxication
Stoned _{(0–100 mm)}	0 [0–0]	1 [0–8]	1 [0–5]
Drug Effect _{(0–100 mm)}	0 [0–0]	2 [0–6]	1 [0–3]
Euphoric _{(0–100 mm)}	0 [0–8]	5 [0–24]	2 [1–16]
Positive Oral Fluid Drug Tests (n)	1	0	1
Measures of Stress ^d
Subjective Stress
Calm _{(0–100 mm)}	76 ± 19	51 ± 24	78 ± 14
Bored _{(0–100 mm)}	18 [7–45]	15 [3–44]	22 [7–60]
Nervous _{(0–100 mm)}	32 [17–59]	27 [5–56]	5 [1–15]
Excited _{(0–100 mm)}	57 [50–67]	26 [14–55]	19 [8–50]
SBP (mmHg)	117 ± 9	122 ± 15	117 ± 11
DBP (mmHg)	78 ± 7	81 ± 12	78 ± 10
Measures of Lipolysis
Plasma Glycerol Concentrations (mg/L)	4.17 ± 3.11	5.73 ± 3.44	5.00 ± 3.00
Plasma FFA Concentrations (mEq/L)	0.28 ± 0.15	0.33 ± 0.16	0.44 ± 0.22

Open in a new tab

Values are frequency (n), mean ± standard deviation, or median [IQR], as appropriate (i.e., where data are normal and non-normal, respectively). DBP: Diastolic Blood Pressure; FFA: Free Fatty Acids; SBP: Systolic Blood Pressure. a: Plasma 11-OH-THC, CBD, 7-COOH-CBD, 7-OH-CBD, 6-OH-CBD, CBG and CBDV concentrations are not presented as these analytes were seldom detected; b: Plasma 7-OH-CBD, 6-OH-CBD, CBG and CBDV concentrations are not presented as these analytes were never detected; c: Oral fluid THC concentration is not presented as THC was only occasionally detected; d: Heart rate and tympanic temperature are reported in Fig. 4

Fig. 2 — The change in (A) Plasma THC, (B) Plasma 11-COOH-THC, (C) Plasma Glycerol and (D) Plasma Free Fatty Acid (FFA) concentrations from pre-intervention (–35 min) to post-intervention (+ 5 min) and 2 h post-intervention (+ 120 min). The shaded region represents the intervention (Cold Water Immersion). Each line represents one individual participant. a: Differs from pre- and post-intervention (p < 0.05); b: Differs from pre-intervention (p < 0.05) (as raw values)

Table 3.

The probability (p) values for the fixed effects included in each statistical analysis

		Fixed Effects
		Time		Sex		TimeSex
	Model	F or χ²	p value	F or χ²	p value	F or χ²	p value
Primary Outcomes:
Plasma THC	Gamma	χ²₂ = 0.103	0.950	-	-	-	-
Plasma 11-COOH-THC	SQRT	F_{[2, 25.0]} = 12.1	< 0.001	F_{[1, 13.0]} = 0.300	0.594	-	-
Secondary Outcomes
DSST Correct Responses	Poisson	χ²₂ = 0.548	0.760	-	-	-	-
DSST Attempts	Poisson	χ²₂ = 0.554	0.758	-	-	-	-
DSST Response Accuracy	Gamma	χ²₂ = 2.19	0.335	-	-	-	-
DAT Hits	Poisson	χ²₂ = 0.531	0.767	-	-	-	-
DAT Response Time	SQRT	F_{[2, 26]} = 0.915	0.413	F_{[1, 13]} = 2.13	0.168	F_{[2, 26]} = 0.778	0.470
DAT Tracking Error	Standard	F_{[2, 26]} = 1.82	0.182	F_{[1, 12]} = 5.27	0.041	-	-
PSAT Correct Responses	Poisson	χ²₂ = 1.14	0.565	-	-	-	-
PSAT Response Time	SQRT	F_{[2, 25.2]} = 0.798	0.461	F_{[1, 13.1]} = 0.448	0.497	F_{[2, 25.2]} = 2.21	0.130
Stoned ^a	Gamma	χ²₂ = 12.9	0.002	-	-	-	-
Drug Effect ^a	LOG	F_{[2, 28]} = 7.72	0.002	-	-	-	-
Euphoric ^a	SQRT	F_{[2, 28]} = 3.06	0.063	-	-	-	-
Measures of Stress
Calm	Gamma	χ²₂ = 11.3	0.003	-	-	-	-
Bored ^a	SQRT	F_{[2, 26]} = 0.221	0.803	F_{[1, 13]} = 0.057	0.815	F_{[2, 26]} = 1.548	0.232
Nervous ^a	SQRT	F_{[2, 26]} = 5.06	0.014	F_{[1, 13]} = 0.993	0.337	F_{[2, 26]} = 0.333	0.720
Excited ^a	SQRT	F_{[2, 28]} = 10.5	< 0.001	-	-	-	-
SBP	SQRT	F_{[2, 28]} = 4.47	0.021	F_{[1, 13]} = 16.1	0.001	-	-
DBP	Standard	F_{[2, 26]} = 1.67	0.208	F_{[1, 13]} = 1.46	0.248	F_{[2, 26]} = 1.02	0.376
HR	SQRT	F_{[6, 80.4]} = 8.58	< 0.001	-	-	-	-
Tympanic Temperature	Gamma	χ²₆ = 25.0	< 0.001	χ²₁ = 2.69	0.101	-	-
Measures of Lipolysis
Plasma Glycerol ^b	Standard	F_{[2, 24.1]} = 5.90	0.008	-	-	-	-
Plasma FFA	LOG	F_{[2, 24.1]} = 8.88	0.001	-	-	-	-

Open in a new tab

-: Not required; DAT: Divided Attention Task; DBP: Diastolic Blood Pressure; DSST: Digit Symbol Substitution Task; FFA: Free Fatty Acids; HR: Heart Rate; LOG: Dependent variable was log-transformed; PSAT: Paced Serial Addition Task; SBP: Systolic Blood Pressure; SQRT: Dependent variable was square root-transformed. a: A small positive constant (half the smallest non-zero value) was added prior to transformation to remove zeros; b: ‘Plate’ (Plate 1 vs. Plate 2) was a fixed effect in this analysis (based on visual inspection of the data) (F_{[1, 11,5]} = 28.0; p < 0.001). Fat Mass and the Time by Fat Mass interaction were not included in any of the models

11-OH-THC, CBD and 7-COOH-CBD were only detected (i.e., > 0.1 ng/mL) in a subset of samples (i.e., < 4 participants), while 6-OH-CBD, 7-OH-CBD, CBG and CBDV were not detected (i.e., > 0.1 ng/mL) in any samples. Hence, these data were not subjected to analysis.