For millennia, there has been great interest in the use of cannabis and its derivatives in the treatment of various medical and neurologic problems, including epilepsy.1, 2 Cannabidiol (CBD), a phytocannabinoid compound derived from the cannabis plant, is of particular interest as a potential anticonvulsant due to the reported lack of psychoactive effects compared to those seen in tetrahydrocannabinol (THC). CBD has long been known to be effective in animal models of epilepsy.3–7 However, until recently, few data were available regarding its safety and efficacy in humans. Recent placebo‐controlled studies indicate that CBD may be effective for the treatment of difficult‐to‐control epilepsies including Dravet and Lennox‐Gastaut syndromes.8–10 These experimental data are supported by observational data from class IV open‐label compassionate‐use CBD state‐sponsored expanded‐access programs, such as the one in Alabama recognized under the name “Carly’s Law.”11–13
CBD modulates several cytochrome P450 (CYP) enzymes, which are of potential interest in investigating interactions with other medications. It is a potent inhibitor of CYP2C19, CYP2D6, and CYP2C9 and a potential inhibitor of the CYP3 family.14–20 Based on what is known about CBD’s metabolism and the metabolism of other antiepileptic drugs (AEDs), one could speculate that there could be many interactions given the pervasive involvement of these CYP enzymes in the metabolism of AEDs.21, 22 To date, there are few data on CBD’s interactions with other AEDs. An animal study using maximal electric shock and audiogenic seizure models showed that CBD potentiated the anticonvulsant effects of phenytoin by twofold, and modestly potentiated the effect of phenobarbital; CBD also reduced the anticonvulsant properties of chlordiazepoxide, clonazepam, and ethosuximide.23 These results were thought to be due to CBD’s actions on the CYP system, although these specific actions were not known at the time. A recently published article revealed a clear drug–drug interaction between CBD and clobazam in a group of 13 pediatric patients. Clobazam and, to a greater extent, N‐desmethylclobazam (the active metabolite of clobazam) levels increased in response to increasing doses of CBD.24 Interactions with other AEDs have not been investigated to date.
The University of Alabama at Birmingham (UAB) CBD open‐label compassionate‐use study is investigating CBD as a potential add‐on therapy for the treatment of treatment‐resistant epilepsy in children and adults. All the enrolled patients were taking multiple concomitant AEDs. As part of the study, frequent monitoring of serum AED levels was performed to identify potential interactions, changes in drug levels, and possible clinically significant relationships between drug level changes and any observed adverse events. Given what is known about CBD’s mechanism of action and metabolism, it was suspected that other AEDs with metabolism hinging on similar enzymes would be affected. The purpose of this study was to identify which AEDs potentially interact with CBD based on serum drug level changes, and in turn, if these potential interactions were clinically meaningful. Based on recently published studies and the known CBD metabolic pathways, we hypothesized that as the dose of CBD increased, the levels of clobazam and N‐desmethylclobazam would increase.20, 24, 25 This potential interaction was hypothesized to contribute to a significant increase in reported excessive sedation in participants taking CBD. Based on observations during the study, we hypothesized that participants taking concomitant valproate with CBD would have a higher incidence of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) abnormalities compared to participants not taking valproate. We also hypothesized that there are interactions with other AEDs given CBD’s known action on key CYP enzymes, and thus we examined data on all AEDs used in the study by the enrolled participants.
The UAB CBD program conducts a prospective compassionate‐use open‐label study in patients with treatment‐resistant epilepsy. “Treatment resistant” was defined as failing to respond to a total of four or more AEDs at adequate dose, including at least one trial of two concomitant AEDs. Other inclusion criteria include age >1 year, stable neurostimulator settings, and ketogenic diet ratio for ≥3 months if applicable, documentation of a detailed seizure diary 3 months prior to enrollment, and current Alabama residency. Exclusion criteria include history of substance abuse or addiction, use of medical marijuana or CBD‐based product within the last 30 days, history of allergies to CBD or marijuana products or to sesame, felbamate therapy initiation within the last 12 months, AST or ALT elevation ≥5 times upper limit of normal, hemoglobin <10, hematocrit <30, or white blood cell count <2,000, among others (all inclusion and exclusion criteria are available at www.uab.edu/cbd). Of importance for these analyses is that all participants must have been taking a stable dose of all AEDs for at least 1 month before they could be considered for enrollment. In addition, baseline plasma levels of all AEDs were obtained prior to enrollment.
All participants were seen and evaluated in a devoted research clinic that was held weekly for both adult and pediatric arms of the study. Each potential participant in the study must have a packet submitted on their behalf by their primary neurologist, and all study packets were reviewed by a study approval committee before a participant was enrolled, with second verification by the study principal investigators. At the time of the enrollment visit, all participants and/or their legal representatives signed the consent form that was approved by the University of Alabama at Birmingham Institutional Review Board. The study is U.S. Food and Drug Administration (FDA) approved and registered with www.clinicaltrials.govund
Upon approval for enrollment, participants were seen in the clinic every 2 weeks during active titration of the CBD dose, with less frequent appointments when doses were not adjusted. All participants received the active study drug. Epidiolex is formulated in sesame oil and administered orally.26 Participants were weighed at every clinic visit. All participants were started on a dose of 5 mg/kg/day split between a.m. and p.m.; participants were instructed to take the CBD at the same time they administer their other AEDs dosed twice daily (although this was not strictly monitored). At each follow‐up clinic visit, the dose could be titrated in 5 mg/kg/day increments to a maximum dose of 50 mg/kg/day. Dose adjustments were based on participants’ response to treatment and tolerability. The dose could be decreased over the phone between clinic visits if there are reports of worsening seizures or side effects, but dose increases were made in person only. Study visit interval was extended once a participant had reached a stable dose of CBD for greater than 2 visits.
At each clinic visit, participants received a full physical examination including neurologic examination and laboratory testing, and their seizure diary and side effect profiles were reviewed. Laboratory testing included complete blood count with differential, comprehensive metabolic panel, and serum AED levels, which were obtained at almost all clinic visits; blood for labs was drawn between 11:00 a.m. and 3:00 p.m. Not all blood samples were collected at trough; therefore, some blood levels were not trough concentrations. Investigators had the authority to adjust the dose of other AEDs if their clinical impression was that a reported adverse event is related to the AED itself or a suspected interaction between that AED and CBD—this follows the design of a naturalistic follow‐up study.27, 28 In some cases, particularly with valproate and clobazam, AEDs were weaned and ultimately discontinued due to reports of certain adverse events, notably sedation and elevation of liver function test results.
This analysis encompasses the first year of data collection. AED levels from 39 adults and 42 children were available for analyses. Prior to enrollment, baseline drug levels were obtained by the referring physician and analyzed locally or by designated reference laboratories. Blood samples for drug levels during the study were all drawn in the UAB Kirklin Clinic or Children’s Hospital of Alabama (COA) laboratory and were analyzed either by the UAB or COA laboratory or reference laboratories, as necessary; if reference laboratories were utilized, the same service was used each time a specific test was obtained. For example, testing of clobazam and desmethylclobazam was always conducted by Mayo Clinic. All testing was performed in Clinical Laboratory Improvement Amendments (CLIA)‐certified laboratories (www.cms.gov/Regulations-and-
In addition to the baseline AED levels obtained before study enrollment, each visit in which AED levels were obtained was considered a data point. Each data point was associated with a corresponding CBD dose. Using a mixed linear model (the Mixed Procedure), the data points associated with each AED were analyzed to investigate if the plasma levels of each AED changed significantly with increasing CBD dose. In this analysis, absolute changes in serum levels were investigated only; there was no threshold set for each drug to determine clinical significance. To further quantify the degree of change in AED levels on CBD and to account for the noise in the data related to the naturalistic study design, the baseline and first two recorded AED levels were plotted by the mean change in level between measurements (CBD dose was not accounted for in this analysis). Separate subanalyses were performed to determine if there was a significant increase in reported sedation in participants who were taking concomitant clobazam, and if there was a significant increase in AST and ALT levels in participants who were taking concomitant valproate compared to baseline liver function test values obtained at study entry. These adverse events did not occur frequently with the other AEDs, and thus due to lack of clinical suspicion, these subanalyses were performed only on participants taking clobazam and valproate.
Demographic and clinical characteristics of the participants are included in Table1. Data were analyzed within individual adult and pediatric arms and combined. In some cases, data for a specific AED could be analyzed only within one arm due to insufficient data points for analysis in the other arm (carbamazepine, eslicarbazepine, ezogabine, and pregabalin in children, and ethosuximide and vigabatrin in adults). However, in several cases there were >100 data points analyzed for each AED (Table1). Analysis via the Mixed Procedure allowed us to control for nonuniform changes in both CBD dose and AED doses as allowed in the naturalistic study design. Linear increases in serum levels of topiramate (p < 0.001), rufinamide (p = 0.004), andN‐desmethylclobazam (p < 0.001, active metabolite of clobazam), and linear decreases in levels of clobazam (p < 0.001) with increasing CBD dose were seen in combined pediatric and adult arms (the decreases in clobazam level were concurrent to the decreases in dose as directed by the presence of the adverse effect—sedation). In addition, a significant increase in serum levels of zonisamide (p = 0.017) and eslicarbazepine (p = 0.039) with increasing CBD dose was seen in the adult arm only. Of note, there were no pediatric participants enrolled in the study who were taking eslicarbazepine at the time of this analysis, and thus an interaction between CBD and eslicarbazepine could not be performed in the pediatric arm. There were no significant changes in drug levels with CBD dose titration in the other AEDs analyzed (valproate, levetiracetam, phenobarbital, clonazepam, phenytoin, carbamazepine, lamotrigine, oxcarbazepine, ethosuximide, vigabatrin, ezogabine, pregabalin, perampanel, and lacosamide). T