Objective: Add‐on cannabidiol (CBD) reduced seizures associated with Dravet syndrome (DS) in two randomized, double‐blind, placebo‐controlled trials: GWPCARE1 Part B (NCT02091375) and GWPCARE2 (NCT02224703). Patients who completed GWPCARE1 Part A (NCT02091206) or Part B, or GWPCARE2, were enrolled in a long‐term open‐label extension trial, GWPCARE5 (NCT02224573). We present an interim analysis of the safety, efficacy, and patient‐reported outcomes from GWPCARE5. Methods: Patients received a pharmaceutical formulation of highly purified CBD in oral solution (100 mg/ml), titrated from 2.5 to 20 mg/kg/day over a 2‐week period, added to their existing medications. Based on response and tolerance, CBD could be reduced or increased to 30 mg/kg/day. Results: Of the 330 patients who completed the original randomized trials, 315 (95%) enrolled in this open‐label extension. Median treatment duration was 444 days (range = 18–1535), with a mean modal dose of 22 mg/kg/day; patients received a median of three concomitant antiseizure medications. Adverse events (AEs) occurred in 97% patients (mild, 23%; moderate, 50%; severe, 25%). Commonly reported AEs were diarrhea (43%), pyrexia (39%), decreased appetite (31%), and somnolence (28%). Twenty‐eight (9%) patients discontinued due to AEs. Sixty‐nine (22%) patients had liver transaminase elevations >3 × upper limit of normal; 84% were on concomitant valproic acid. In patients from GWPCARE1 Part B and GWPCARE2, the median reduction from baseline in monthly seizure frequency assessed in 12‐week periods up to Week 156 was 45%–74% for convulsive seizures and 49%–84% for total seizures. Across all visit windows, ≥83% patients/caregivers completing a Subject/Caregiver Global Impression of Change scale reported improvement in overall condition. Significance: We show that long‐term CBD treatment had an acceptable safety profile and led to sustained, clinically meaningful reductions in seizure frequency in patients with treatment‐resistant DS. [ABSTRACT FROM AUTHOR]