Abstracts

Rationale: Cannabidiol has shown utility in patients with treatment-resistant epilepsy, however data regarding the impact of CBD on the EEG is scarce. We assess the longitudinal impact of plant-derived highly purified CBD (Epidiolex®;100mg/ml oral solution) on the EEG of children and adults participating in a state-sponsored, open-label compassionate use study. Patients with video-EEG confirmed treatment-refractory epilepsy who have failed at least 4 different AEDS and with on average at least 4 countable seizures per month are included. Methods: Patients with video-EEG confirmed treatment-refractory epilepsy who have failed at least 4 different AEDS and with on average at least 4 countable seizures per month are eligible for inclusion in the UAB CBD Program.  All participants received least 20-minute long EEG prior to starting CBD. CBD was initiated at 5mg/kg/day and titrated upwards for optimal seizure control every 2 weeks by 5mg/kg/day. Follow-up EEG was obtained at 2 weeks following stable dose of 25 mg/kg/day (approximately 12 weeks into the study). Titration of CBD was continued in selected persons up to the maximum dose of 50 mg/kg/day, with repeat EEG obtained after one year. Of 169 participants enrolled, 122 (60 adult and 62 children) completed first on-drug EEG (follow up 1/FU1), while 67 (40 adult and 27 children) completed a second on-drug EEG (FU2). Pre-/on-CBD EEGs were independently reviewed by two investigators. Reviews were compared and discrepancies mediated. Data were analyzed for the following criteria: background frequency, focal slowing, reactivity, frequency of interictal epileptiform discharges (IED), and total number of seizures. For comparison, each patient served as his/her own control.  For each of the binary outcomes, the longitudinal generalized estimating equations (GEE) for logistic regression was used to regress the logit of the probability of the outcome of interest on last CBD dose: time (treated categorical: baseline, FU1, FU2), patient groups (Pediatric, Adult), and the interaction of time and patient groups to assess longitudinal changes in the likelihood of the outcome variable across patient groups. A linear mixed effects model with random intercepts was used with the same covariates to analyze background frequency. For IED per unit time, the GEE with a specified negative binomial distribution was used to analyze counts of IED with the same covariates as in the previous model but with duration of IED serving as an offset variable in the model. Estimated mean changes, odds ratios and incidence density ratios were used to assess longitudinal effects on outcome variables across patient groups. An alpha level of 0.05 was used as cutoff for statistical significance. Results: The contrast test for time and patient groups interaction in the GEE model for IED frequency yielded statistically significant (P<0.0001) implying longitudinal changes in IED per unit time that depended on patient groups. Specifically, at FU1 relative to baseline the adult group showed significant decreases in IED/minute (IDR 0.07, 95% CI [0.04, 0.14], P<0.001) while a nonsignificant decrease was observed among children (IDR 0.87, 95% CI [0.47, 0.64], P=0.67). The difference in changes (IDR 11.8, 95% CI [4.86, 28.65], P<0.0001) between patient groups at FU1 was significant after adjusting for last CBD dose. At FU2 both groups showed significant decreases from baseline after controlling for last CBD dose, more pronounced in the pediatric group (IDR 15.38, 95% CI [4.93, 47.99], P<0.001). Evidence from data analysis was insufficient to support an association between last administered CBD dose and longitudinal changes in IED. Conclusions: In this study, we show a longitudinal reduction in IED frequency in patients with difficult to control epilepsies treated with CBD. Funding: The State of Alabama and Carly's Law