When cannabis was originally scheduled by the U.S. Drug Enforcement Agency (DEA) and the Food and Drug Administration in 1970, its procurement, use, and sale were heavily associated with illicit use [1]. Cannabis was first legalized for restricted medical use by the state of Colorado 30 years later, in 2000. Since then, U.S. state-level cannabis legality has boomed, with 38 states legalizing cannabis for medical use, 23 states legalizing recreational use, and 9 states starting ‘low tetrahydrocannabinol (THC)’ programs. As of March 2024, cannabis was recommended to be federally re-scheduled from a Schedule I drug to a Schedule III drug [2].

Despite this widespread legalization in the U.S., regulation of commercially available cannabis products is not universal, with variation between states regarding available products, labeling requirements, and consumer education [3]. Because of its historical, federally illegal status, cannabis has not been regulated by the Food and Drug Administration (FDA). Although the FDA has approved one cannabis-derived drug and three cannabis-related drug products, the agency has not approved a “marketing application” for cannabis to treat illness and does not regulate the drug [4]. The FDA regulates label requirements of legal drugs, and in lieu of federal regulations, cannabis label policies fall to the discretion of the states [5]. The evolution of THC concentration testing and labeling regulations has lagged behind the rapidly growing market and product types, and there are perpetual concerns about product regulation and safety [6]. Of concern to state regulators and consumers alike is the accuracy of product labels, including the concentration of THC, the primary psychoactive component.

Label accuracy is an element of product regulation that refers to how well a label represents the content of the product (i.e., accurate representation of THC content). A 2022 study found that there is a lack of uniform labeling practices across states with legal cannabis [5]. The same study found that, on average, states have 16.7 different required label attributes and that there is no relationship between years of legal status and labeling requirements. Uniformity in labeling guidelines (as with food labels) promotes awareness of product content and how labels may be interpreted.

One aspect of labeling is THC concentration and the associated permitted variability in concentration. In Colorado, the regulations for label claims for any cannabinoid that has a 2.5 milligram serving or more may be within plus or minus 15% of the concentration actually contained in the product (i.e., cannabis flower that is labeled as 30% THC may actually contain 25.5% to 34.5%). Products with less than 2.5 milligrams per serving may be within +/- 40% of what is reported on the label [7]. These regulations were based on plausible variations in concentration, with consideration of other relevant industries and laboratory testing procedures. Other states’ regulations on legally allowable concentration variation are informative. For example, the state of California requires labels to be within +/- 10% of the observed THC concentration [8,9]. Massachusetts’s law states that cannabis labels are required to display “the amount of delta-nine-tetrahydrocannabinol (∆9-THC) in the package and in each serving” of a marijuana product as expressed in absolute terms and as a percentage of volume [10].

Limited empirical research has reported independent test results of the THC concentration in products available in the commercial marketplace. In a study of the accuracy of cannabis flower labels in Colorado, researchers tested a sample of N = 23 different products and found that 78% of the products were over-labeled (i.e., labeled as having an average of 15% higher THC concentration than was found in independent testing) [6]. In a study of label accuracy of cannabis edibles from California and Washington, Vandrey et al. also found that 60% of products in their study were labeled as having higher THC content than was observed in the product [11]. These studies show that product labels may consistently inflate the actual THC concentration in the product [6,11]. This is important because data also show that people rely on reported concentration for purchasing and dosing decisions [12]. A major gap in these a priori studies is the absence of testing extracted or concentrate products, including edibles, oils, and concentrates (e.g., dabs, shatter, wax, resin), which make up 30% of the legal market [13,14].

The aim of this study was to determine the accuracy of cannabis product labels with respect to THC content among a diverse set of samples of cannabis products (e.g., flower, edibles, and concentrates) purchased in Colorado retail or medical dispensaries as part of a larger study. We explored THC accuracy in the context of product labels and Colorado state regulations. Further, we explore differences between in accuracy by product type.

This study independently tested a range of cannabis products purchased from cannabis dispensaries in Colorado. This study is consistent with the limited studies that have examined this, which have been conducted in several different states, that find cannabis flower and edible products may not be accurately labeled and do not always reflect observed THC concentrations [6,11]. Corroboration of previous findings is significant and suggests that THC concentration in cannabis products is inaccurately marketed across a number of products. Our findings extend prior research by including extracted and concentrate products (e.g., edibles, vape cartridges, shatter) and included edibles that are reflective of the current market, which are widely available, and make up a large component of the market [15,16].

Cannabis concentrate products are sometimes purported to have risk of greater public health concerns, such as the potential for greater adverse events associated with higher dose [17], perhaps an unexpected positive side effect of the extraction process is the potential for greater accuracy of product labeling. Although our study found some concentrate products (n = 4) were not labeled within Colorado state regulations, a higher proportion of this category of cannabis products were within the margin of allowable variation.

Label accuracy may be a particularly relevant concern for individuals using cannabis for medical purposes, in which the accuracy of labeling to produce desired and therapeutic effects is a driving motivation for choice of product. There is evidence that people who use cannabis for medical symptom management may purchase products from the recreational or the medical market [18]. These individuals may purchase from either retail or medical sources when both options are available in their state, and thus, product label accuracy is a priority regardless of medical or retail status [18]. This should be of interest to policymakers because medical cannabis use is prevalent in the United States due to the fact that an estimated 3.8 million people are registered as medical cannabis patients [19].

There are significant challenges to product testing, particularly with flower products. The inherent variability in a plant means there will be variability within and between flower buds produced by a plant and its clone. In Colorado, state regulations dictate how flower buds are sampled (e.g., 5 samples must be taken totaling 2.5 grams of total mass for harvest batches smaller than one pound) and that these samples are combined and tested as a single test sample representative of the entire pound of material [20].

There are no universal guidelines for cannabis labeling requirements and variability in state testing and label reporting. This is of concern based on consumer and regulator interest in THC concentrations. We found that across product types (flower, edibles, concentrates), there were differences between labeled and lab-tested THC concentrations, with a pattern that packages reported higher THC levels than were found in products. This systematic pattern of over-labeling cannabis products found by our study and others suggest that differences between product labels and observed THC may not just be biological variability. Rather, they reflect systematic forces and incentives in the marketplace, which lead to over-labeling. With rescheduling on the horizon, it is imperative to strategize ways to incentivize federal regulators (like the FDA) to create cannabis label requirements. In lieu of federal regulation, it is necessary to engage state-level regulators to facilitate the evolution of policy that requires cannabis labels to be reliable and accurate.