The primary findings of the current study were (1) THC T_max ranged between 35 and 90 min, and THC C_max ranged between 3.2 and 5.5 ng/mL. (2) Differences between products in plasma THC concentration during the first 20–30 min were detected. There were no differences between products in the metabolite-to-parent C_max and AUC ratios for THC-COOH and THC-OH. (3) Significant relations were identified between body composition and pharmacokinetic parameters for some products, however none of these body composition characteristics were consistently related to pharmacokinetic parameters across all five of the products. (4) Compared with a control condition, edible marijuana had no effect on any of the indices of glucose regulation, including Matsuda Index, and 2.0 h and 3.5 h AUC.

Commercially available edible marijuana products are becoming increasingly popular alternatives to inhaled combusted marijuana, in part due to the perceived lower cardiopulmonary risk associated with protecting lungs from exposure to smoke. However, as with marijuana prepared for burning/smoking, considerable pharmacokinetic variability exists between edible marijuana products, even when standardized for THC content. For example, in two recent studies of cannabis-infused brownies containing 50 mg of THC, THC C_max ranged between 2.5 ng/mL and 4 ng/mL, and T_max between 1 and 2 h [9,10]. In contrast, in the present study research participants self-administered commercially available edible marijuana containing only 10 mg of THC, and C_max ranged between 3.2 and 5.5 ng/mL, while T_max ranged between 35 and 90 min. Thus, several of the commercially available preparations containing only 20% of the THC content of the brownies evoked greater circulating concentrations of THC, and in a shorter period. To account for these discrepancies between studies, we have considered several potential explanations. In addition to cannabis, the brownies contained additional ingredients, including fats and sugars, that contributed to a greater caloric load compared with a single (10 mg THC) serving of any of the commercial edible marijuana studied in the current investigation; we conservatively estimate 100 vs. 0–30 kcal, respectively. It is plausible that the additional caloric load of the brownies interfered with the rate and amount of THC absorption from the gut [29]. Second, the brownies were typical/normal food supplemented with cannabis, whereas the commercial products under investigation were engineered to provide superior delivery of THC. This engineering includes the development of water-soluble THC and marijuana concentrate, and controlled addition of solubility agents, such as gum arabic and medium-chain triglycerides (MCTs), that have been shown to promote delivery and bioavailability of dietary supplements [30,31]. Third, in a separate brownie study [8], approximately 50 mg of THC incorporated within a brownie produced a THC C_max (measured in whole blood) of 4.7–34.8 ng/mL in frequent marijuana users (≥5 times per week) and 3.6–22.5 ng/mL in occasional marijuana users (≤3 times per week). In the same study, THC T_max was 1.5–3.5 h for both frequent and occasional users. These data imply that the degree of habitual marijuana use may promote THC C_max independent of T_max, thus when comparing across studies, data must be interpreted with caution if habitual marijuana use is not reported. In the current study, a repeated measures crossover design was employed, thus comparison within the study should not be compromised by heterozygous habitual cannabis/marijuana use; this is a strength of our study.

Previously, it has been reported that parameters of oral CBD pharmacokinetics, including T_max, were associated with anthropometric variables, including fat-free mass and body mass index [11]. A number of explanations were proposed, such as the positive relationship between fat-free mass and blood volume [32,33], representing a greater reservoir in which to dilute circulating metabolites; greater perfusion of metabolically active tissues, thereby leading to enhanced clearance of circulating metabolites; and the lipophilic nature of CBD. In the current study we extend these observations and arguments to consider a potential role of body composition specific to THC pharmacokinetics. To our knowledge, our data are the first to directly address this issue. While significant relations were identified between body composition and pharmacokinetic parameters for some of the edible marijuana products (Table 7), none of these body composition characteristics were consistently related to pharmacokinetic parameters across all five of the studied products. Similar to CBD, THC is thought to be lipophilic and consequently absorbed by adipose tissue. Two lines of evidence from frequent marijuana users support this claim: THC is quantifiable in fat biopsies taken 4 weeks after previous marijuana use [12] and circulating THC concentration is increased following 35 min of lipolysis-inducing moderate intensity stationary cycle ergometer exercise [34]. Similar exercise-mediated observations have been reported in exercising rats pre-treated with THC [35], but not reproduced in a smaller follow-up study of humans with low-to-recommended body mass index (BMI; range: 19–23 kg/m²) [36]. Our analysis revealed inconsistent correlations for absolute and/or % fat mass for T_max, C_max and AUC_0-240 for only some of the products. Although our study population was small, the range in adiposity and BMI was large, ranging from lean to obese. The heterozygous nature of the body composition data should have facilitated the identification of potentially significant correlations. While it appears likely that in vivo THC is lipophilic and absorbed by adipose tissue, adipose may not appreciably contribute to short-term edible marijuana pharmacokinetics.

The final aim of the current study was to explore the potential acute influence of edible marijuana on glucose regulation. The connection between marijuana use and the development of insulin resistance and diabetes is not clearly understood. On one hand, there exists an intuitive link: marijuana use is associated with increased appetite [37], leading to increased dietary intake, and in particular increased purchase of foods of high caloric but poor nutritional value (“junk food”) [38]. If this behavior is chronic then weight (fat) gain typically ensues, and the overweight and obese states are causally tied to insulin resistance and diabetes. However, on the other hand, most of the epidemiological research suggests that marijuana use either has no effect on diabetes risk, or in many cases lowers the risk of diabetes, even when taking into account other potential confounding variables such as body composition [19,20,21,22,23]. The collective concluding recommendation provided by the authors of these studies was the need for randomized controlled trials to provide definitive insight. To our knowledge, aside from some preliminary work completed in the 1970s [24,25,26], there have been no human intervention studies to support the epidemiology. The early preliminary work involved intravenous marijuana administration and inhalation of combusted marijuana, and resulted in either impaired or unaltered acute glucose regulation. Inhalation of combustible materials typically leads to temporarily decreased oxygen transport (i.e., relative hypoxia) and increased sympathoadrenal activity, two conditions known to unfavorably modify glucose control [39,40]. Further, since these early studies, the THC potency of marijuana products has increased [27] and the methods of marijuana consumption have become more diverse [1], thus our goal in the present study was to update the paradigm using commercially available edible marijuana with THC doses reflective of contemporary use. The gold standard measurement of glucose regulation and insulin sensitivity is the hyperinsulinemic euglycemic clamp technique, involving the intravenous co-administration of insulin and glucose. Unfortunately, on account of state laws and institutional regulations, marijuana is not permitted in our university clinical research facility, thus we compromised by studying glucose regulation using a modified OGTT in a field setting. None of the five commercially available edible marijuana products influenced the acute circulating glucose response to standardized glucose ingestion (p = 0.98). The variability in circulating glucose across conditions was akin to desirable studies of test–retest reliability [41]. The primary hormone involved in glucose regulation is insulin. Improvements in glucose control are not always necessarily reflected by lower circulating glucose concentrations [42]; improved glucose regulation may be reflected by lower insulin in the absence of modified glucose values. For this reason, we compared the OGTT insulin response in the control condition to the OGTT insulin response to two of the five edible marijuana products. There were no differences in circulating insulin between products, nor were there any differences in glucose/insulin derived indices of glucose regulation, including the Matsuda Index. Based on our data, and in light of the epidemiological studies [19,20,21,22,23], we suggest that, in adults free from diabetes, edible marijuana does not affect acute glucose regulation, but potentially may exert a favorable chronic influence via long-term physiological responses such as decreased adiposity [43], and improved gut [44] and/or liver health [45]. Noteworthy, recent data from a Mendelian randomization study suggest that chronic use of products derived from Cannabis sativa L. does not influence the prevalence of type 2 diabetes [46]. Clearly the potential connection between marijuana and glucose regulation remains unresolved.

There are several issues and potential study limitations that warrant further discussion. Marijuana and other Cannabis sativa L. products have been purported to have multiple medicinal benefits; the evidence, or lack thereof, to support some of these claims has recently been reviewed [13,14,47,48]. Collectively, these reviews highlight some of the inconsistencies between previous reports and conclude with recommendations for future research and the need for randomized controlled trials. In this regard, marijuana research is difficult. Among some of the problems unique to marijuana research are significant legal and regulatory compliance issues, difficulties associated with research participant compliance pertaining to short-term marijuana abstention, and ethical and safety concerns pertaining to the consumption of psychoactive substances. In addition, specific to marijuana research in the USA, there is currently a restricted number of federally approved cannabis growing facilities, thereby limiting the diversity of strains of plants with varied cannabinoid ratios and potencies available for study, and the requirement of a federal license and registration with the Drug Enforcement Administration for the dispensing of marijuana to research participants. With reference to the current investigation, marijuana is not permitted in our university clinical research facility and investigators are not permitted to handle marijuana. Study limitations resulting from these regulations included the need to transfer research location from a clinical facility to a field location, and compromise with the choice of procedures (e.g., the hyperinsulinemeic euglycemic clamp technique vs. the OGTT). With respect to the former, several of the phlebotomy problems that contributed to participant withdrawal may have been avoided, and the clinical furniture (e.g., adjustable hospital bed) may have promoted participant comfort during the 5 h protocols. As for assessment of glucose regulation, although not the gold standard protocol, the OGTT is generally considered an acceptable alternative to the clamp technique [49]. Additional study limitations include the number of participants who completed the whole study (7/15). While relatively small, the sample size was sufficiently large to allow for detection of differences in circulating THC concentration across marijuana products, and for the identification of significant associations between pharmacokinetic parameters and body composition variables. We did not detect an influence on glucose regulation between any of the marijuana products and the control products; while it is feasible that this could be attributed to a type 2 error, power analysis [50] suggested that based on our current data, and assuming a similar degree of variability, approximately 200 participants would have been required to detect an influence of marijuana on the Matsuda Index. Noteworthy, we studied young adults without diabetes. It is plausible that marijuana may have had a more pronounced effect on glucose regulation in adults with poor glucose control, including those with diabetes, or adults at risk for diabetes, such as older adults [51], and/or sedentary adults with body mass index values greater than 25 (overweight) or 30 kg/m² (obese) [52]. In light of the difficulties associated with marijuana research, an alternative approach could be to persuade regular users of marijuana to abstain for a brief period, and quantify glucose regulation before and after abstention, and again after resumption of normal use. The hypothesis would be that glucose regulation was unfavorably influenced by marijuana abstention. Such an approach would circumvent the need to administer marijuana, but compliance with abstention, confirmation of compliance, and the substitution of marijuana for another behavior may introduce new challenges and/or confounding variables.

Limitations aside, we believe our data to be novel and of practical value. To our knowledge, our study represents the first to compare the pharmacokinetic properties of commercially available edible marijuana products and to evaluate the potential influence of body composition. Further, our data may be used by investigators to plan the timing of measurements and procedures in future studies exploring the physiological effects of edible marijuana.

In summary, in recreational users, knowledge of pharmacokinetics associated with different commercially available edible marijuana products may prevent accidental overdose due to premature repeat dosing. THC T_max and C_max ranged between 35 and 90 min, and between 3.2 and 5.5 ng/mL, respectively. Differences exist between commercial products with respect to plasma THC concentration during the first 20–30 min following ingestion. Body composition does not appear to be an acute determinant of edible marijuana pharmacokinetics. Edible marijuana does not appear to affect acute glucose regulation, but in light of epidemiological data, marijuana potentially may exert a favorable influence on long-term glucose regulation and diabetes risk.