PND20 COST-UTILITY OF CANNABIDIOL IN ADDITION TO BACKGROUND THERAPY VERSUS BACKGROUND THERAPY ALONE IN LENNOX-GASTAUT SYNDROME

Abstract

Lennox-Gastaut Syndrome (LGS) is a rare, severe, difficult to manage form of epileptic encephalopathy, for which cannabidiol (CBD) was recently approved. This analysis assessed the lifetime cost-effectiveness of daily, oral CBD 10 mg/kg add-on therapy compared to standard of care in these patients. We developed a Markov decision analytic model using inputs from the GWPCARE3 phase III clinical trial, and other published literature. Patients in the base case analysis represented the clinical trial population. Health states represented a quartile percentage reduction of “drop attacks” from baseline frequency. Seizure-related costs in 2018 USD and quality-adjusted life years (QALYs) were projected over a lifetime and we used a standard annual discount rate (3%). For the base case, we assumed treatment effect persisted throughout the lifetime. Overall trial results were applied to the individual age group entering the model, and patients did not discontinue therapy. Scenario analyses were conducted to test various assumptions. One-way sensitivity analyses were conducted using confidence interval ranges found in the literature, otherwise we used a default of 25%. A probabilistic sensitivity analysis was conducted, and cost-effectiveness acceptability curves were developed. Patients on CBD therapy achieved 1.12 additional QALYs over the lifetime, compared to standard care, and cost an additional $34,269 per person. CBD arm pharmaceutical costs were an additional $66,503, while seizure-related costs were $32,234 fewer. The base case ICER value was $30,573 per additional QALY gained. The addition of CBD to background therapy for patients with LGS appears to be economically feasible, and should be considered a cost-effective therapeutic option. Limitations arise from the scarcity of available data in LGS patients, particularly regarding long-term mortality and duration of treatment effect. Future studies clarifying these data are needed to improve model accuracy and generalizability.