The medical cannabis programs of 34 jurisdictions  varied greatly in their listing of qualifying conditions

Many patients also opt for medical cannabis, which can be easier to access than prescription drugs and has been legalized in more than half of the states in the U.S. . However, medical cannabis has not undergone the U.S. FDA approval process, and is not under the same supply chain controls as other prescribed pharmaceuticals. With the increase in popularity of cannabis and cannabis‐derived products, more attention is given to toxicology and human health risk of cannabis contaminants . Several cannabis product recalls have been issued in the U.S. due to contamination of insecticides  and fungicides. Additionally, there are reports of pesticide spiking in illegal synthetic products, including brodifacoum  and paraquat. Pesticide use in agricultural commodities is regulated under the Federal Insecticide, Fungicide and Rodenticide Act. Yet, due to the federal status of cannabis as a Schedule I substance , the U.S. Environmental Protection Agency  has not issued any guideline on pesticide applications in cannabis. Following the wave of legalization of medical or recreational cannabis across the U.S., there is an expectation of the general public that cannabis legalization also results in regulation to ensure safety in cannabis consumption . In many states, cannabis is recommended by physicians for therapeutic use in various medical conditions. At the same time, there are no federal regulations in place to standardize cannabis as a pharmaceutical. The potential for contamination of cannabis with pesticides is an area of ongoing analysis , and has been observed in medical cannabis samples . The inconsistent regulation of medical cannabis, together with potential exposure to harmful pesticides, can result in adverse health outcomes in patients with susceptible conditions. Here, we examine the state‐level regulations, publicly available pesticide residue testing reports, and curated biological interactions in the Comparative Toxicogenomics Database  to evaluate the potential neurological hazards of pesticide exposure in medical cannabis.

We surveyed the online information provided by the public health agencies and agriculture departments of 50 states and Washington, D. C. between September 15 and November 29, 2020. We first determined whether medical and/or recreational cannabis was legalized in each jurisdiction. If medical cannabis was found legal in a jurisdiction, we would categorize the qualifying conditions with reference to the 2017 National Research Council report, “The Health Effects of Cannabis and Cannabinoids”, which described 21 cannabis treatable diseases with different levels of therapeutic evidence . An earlier study took a similar approach to evaluate the prevalence of qualifying conditions in the U.S. . Here, we mainly focused on neurological diseases in our analysis. We next compared the action levels published by each jurisdiction to regulate pesticide residues in cannabis. If no action level was published online,trim tray we would submit a direct inquiry to the cannabis program. We also checked with ISO/IEC 17025‐ certified laboratories in the state with legalized cannabis . With the passage of the 2018 Farm Bill, pesticide applications in hemp are now regulated by the U.S. Department of Agriculture  under FIFRA . Thus, we excluded the states that only allowed the use of cannabidiol oil in our analysis.We evaluated the potential connections between insecticides, cannabinoids, and seizure using CTD . We searched CTD for specific insecticides and cannabinoids to build sets of computational constructed information blocks  that related a chemical‐gene interaction with a phenotype and seizure, following the methodology previously described . Briefly, five independently curated data sets  were integrated and used as lines of supporting evidence to connect and computationally construct CGPD‐tetramers. Each CGPD‐tetramer represented a potential chemical‐to‐seizure connection that met all five lines of evidence. We also compared the gene connections of the insecticide and cannabinoid CGPD‐tetramers to the 38 gene variants listed in the 2016 and 2018 reports of the International League Against Epilepsy Genetics Commission .We calculated the medians and ranges of pesticide action levels in different jurisdictions. We compared those figures with the tolerances  set for food commodities by the U.S. EPA  and the reported values of pesticide residues in cannabis.

Using Tableau Desktop , we created layered plots that encoded the range of the action levels as gray horizontal lines, and plotted key values as colored circles. In the first chart, the lines served as paths between two values: the minimum and maximum action levels set by each jurisdiction in our data collection. The second chart used a “barbell” style plot, where horizontal lines also served as paths, but these paths connected two different values: the lowest U.S. EPA tolerance levels for food commodities and the median of the action levels. The third chart showed the highest reported values of pesticide residues in cannabis from an open literature search. The action levels, tolerances, and reported values were plotted on a log scale. Using the CTD CGPD‐tetramers, we produced a list of relationships between chemicals and genes, with each relationship weighted by the number of tetramers in the database mentioning the interaction between a chemical and a gene. This produced a weighted edge list that we passed into Gephi, a network analysis and visualization application . Using Gephi, we calculated weighted degree centrality, and used the biological functions of genes as node categories. The result was a bimodal network of chemicals × genes, with each gene and their connections to the chemicals color‐coded by the gene’s biological function. Functional annotation of the genes used the NIH/NIAID  Database for Annotation, Visualization and Integrated Discovery  version 6.8 . Nodes and edges are sized by weighted degree centrality. Larger nodes indicate chemicals and genes that receive more attention in the CTD curated literature.We began by surveying the status of cannabis legalization in 50 states and Washington, D.C. Thirty‐four states and D.C. permitted cannabis use for medical purpose. Since South Dakota legalized both medical and recreational cannabis on November 3, 2020 , the qualifying conditions for medical use were not yet available. The other 16 states allowed the use of cannabidiol oil only.Three of the jurisdictions had specialized programs for adults and a separate restricted list of qualifying conditions for pediatric use of medical cannabis. Three jurisdictions did not list any explicit condition to qualify medical use. Ten jurisdictions gave physicians full discretion to prescribe outside of the listed conditions. Another 11 jurisdictions allowed petitioning on a case‐by‐case basis or adding a new qualifying condition at any time, and two jurisdictions allowed public petitioning during legislation changes.

Table 1 shows a total of 56 qualifying conditions related to neurological dysfunction , psychological conditions , and pain and injuries  as listed by 31 jurisdictions as well as 2 conditions listed in the NRC report that no jurisdictions explicitly allowed. The average number of enumerated conditions per jurisdiction was 17. One jurisdiction listed as broadly as 53 conditions, while another listed only 9 for non‐pediatric prescription. Some of the qualifying conditions were described in language with limited specificity. For instance, many jurisdictions listed “Multiple Sclerosis” as a qualifying condition, but the majority also listed “Severe or Persistent Muscle Spasms”, often in the same sentence. Multiple sclerosis was mentioned together with muscle spasms in 15 jurisdictions. It was mentioned alone in 10 jurisdictions. An additional two jurisdictions listed muscle spasms as a qualifying condition without mentioning multiple sclerosis. Depression and schizophrenia – both of which were reviewed in the 2017 National Research Council report  – were not listed by any jurisdiction. We next examined the listing of 11 neurological categories across the 31 jurisdictions. “Movement Disorders” was the most common neurological category and all 31 jurisdictions listed at least one movement disorder as a qualifying condition . These conditions included epilepsy, certain symptoms of multiple sclerosis, Parkinson’s Disease, mobile vertical rack and any cause of symptoms leading to seizures or spasticity. This was consistent with earlier reports that epilepsy and seizure disorders were the two common conditions qualified for medical use in the U.S. . Based on the language used by these 31 jurisdictions, the authorized use of medical cannabis appeared to be intended to address the movement related symptoms rather than the etiologies of the disorders. “Pain‐Related Conditions” was the second most common category , followed by “Anorexia and Weight Loss”  and “Psychiatric Conditions” . Many of the qualifying conditions were comorbid such as cachexia/wasting syndrome and HIV/AIDS, cancer, or other causes of majorweight loss. Notably, 46 conditions were qualified for medical use by just one jurisdiction .Medical cannabis is a potential route of pesticide exposure to patients with neurological diseases. Instead of alleviating a patient’s condition, the use of cannabis may harm the patient if it is contaminated by pesticides. We investigated the pesticide testing requirement of cannabis in the state‐level jurisdictions with legalized medical use. We found that 24 states and D.C. were posting the pesticide testing requirements and action levels online. We contacted the cannabis programs in the remaining nine states and found that pesticide testing was not required in three states. Also, three states provided no clear response to our inquiries. By the end of this study, we were able to obtain the action levels in 27 states and D.C. In all 28 jurisdictions, pesticide testing of cannabis was required at both the raw agricultural commodity level and the final product level.

Six states – Connecticut, Illinois, Louisiana, Maine, North Dakota, and Ohio – adopted the U.S. EPA tolerances for food commodities as the action levels of pesticide residues in cannabis . In these states, a cannabis sample would pass the pesticide residue test if it satisfied the most stringent tolerance levels for up to 400 pesticides. Maine also banned the use of 195 pesticides in cannabis that were federally prohibited for use on organic produce . Minnesota adopted the pesticide testing guideline for articles of botanical origin provided by the U.S. Pharmacopeia Convention . Twenty states and D.C. took a different approach to assess each pesticide and develop action levels individually.Pesticide exposure can result in adverse neurological effects in humans. For instance, acute poisoning of organophosphate and carbamate insecticides results in cholinergic symptoms . We reviewed the 155 pesticides regulated by the 20 states and D.C. . Insecticides  and fungicides  were the most two regulated classes of pesticides, followed by plant growth regulators , herbicides , and rodenticides . These 155 pesticides also included 16 organophosphate and 8 N‐methyl carbamate insecticides listed in the 2006 and 2007 U.S. EPA reports on cumulative risk assessment . The large number of insecticides and fungicides under regulation reflected the industrial practice of using chemical measures to control mite infestation and powdery mildew . Most of these 21 jurisdictions had action levels for 40–60 pesticides. Abamectin, bifenazate, etoxazole, and imidacloprid were regulated by 20 of the 21 jurisdictions. These four pesticides were also regulated by the six states that adapted the U.S. EPA tolerances. In contrast, 84 pesticides were regulated in only one of the 21 jurisdictions with specified action levels, and only 17 of those were also covered by the U.S. EPA tolerances for food commodities. Lastly, the 155 pesticides regulated by the 20 states and D.C. did not include a number of pesticides previously found in illegal samples, such as brodifacoum, naphthalene, and paraquat . Fig. 3 shows the top 50 pesticides with the largest variation of action levels in 20 states and D.C. On average, the action levels of these 50 pesticides were 32‐fold higher than the most stringent tolerances for food commodities by the U.S. EPA . Sixteen out of the 17 reported values of pesticide residues in cannabis plant matter were above the U.S. EPA tolerances for food commodities . Dimethomorph, a fungicide, showed the largest variation in the action levels, ranging from 0.1 to 60 ppm in 5 states. Azoxystrobin  and chlorantraniliprole  both showed a 4,000‐ fold difference in action levels. The action levels of these two pesticides ranged from 0.01 to 40 ppm in 17 and 12 jurisdictions, respectively. Ethephon, a plant growth regulator, was regulated by nine states for applications in cannabis. Six of these nine states adopted the U.S. EPA tolerance at 0.002 ppm . Two states set their action levels at 1 ppm. The remaining state set its action level at 0 ppm  with a target limit of quantitation of 0.005 ppm. In this state, the laboratories were required to detect at least 0.005 ppm of ethephon using their analytical instrument. If their instrument allowed them to detect smaller quantities of ethephon, any amount detected would cause the sample to fail the testing process.