Rationale: Three large placebo-controlled trials have demonstrated efficacy of cannabidiol (CBD) for refractory epilepsy, whereas data regarding the efficacy of tetrahydrocannabinol (THC) remain scarce. Most published reports on the efficacy of CBD/THC combinations derive from state medical marijuana programs with widespread variation in formulations and minimal physician involvement. In contrast, the Texas Compassionate Use Program (CUP) includes a centralized database, requires physician oversight of therapy, imposes maximal THC concentration and limits production to only three licensed dispensaries. These features grant unique opportunities to monitor clinical response and tolerability with consistency in CBD:THC products used throughout the duration of therapy. Methods: This is an IRB-approved, unblinded, retrospective cohort study of patients with drug-resistant epilepsy (DRE). Subjects enrolled in the CUP and prescribed cannabis-derived CBD/THC preparations from a single dispensary source by two neurologists, who adjusted dosage without pre-specified endpoints, were later consented for review of their clinical outcomes. Record review from the first clinic visit after starting treatment (minimum eight weeks later) determined percent seizure reduction as the initial primary outcome, and therapy changes were traced over subsequent visits to find enduring seizure response, reason for discontinuing, maximum tolerated dose, side effects, impact of crossover to pure CBD (Epidiolex®), and duration of CBD therapy. Results: Of 212 patients enrolled and prescribed CBD/THC, 135 subjects with DRE were consented and 117 met criteria with follow-up data sufficient for analysis. Patients continued CBD therapy for a mean of 429 days (range 7-868). Thirteen percent reported 50-90% seizure reduction and 21% reported >90% reduction (Figure 1). Consistent or further improved seizure reduction continued for 91% of all patients over the course of follow-up, as reported at last clinic follow-up (Figure 2). Ninety-one patients remained on treatment, 58% of whom continued a combination of Epidiolex® and dispensary CBD/THC. Reported adverse effects included somnolence, mood/behavior changes, increased seizure frequency, and gait disturbance. Common reasons for discontinuing therapy included cost and lack of efficacy. Some families continued therapy to preserve unexpected benefits such as calmer behavior or improved sleep, in some cases despite little change in seizure activity. Conclusions: The size of our cohort, single source of stable combination therapy, and duration of follow-up significantly expand the outcome data of CBD/THC therapy for epilepsy. Results suggest that initial response to CBD can inform enduring response to therapy, in part reflecting a selection bias of retrospective review. Many patients/families perceived benefits beyond seizure control by continuing dispensary CBD after starting Epidiolex®. There is further need for prospective study to evaluate the potential of CBD/THC combination therapy. Safety concerns regarding cognitive impact and risk of psychosis ascribed to THC exposure demand better understanding in therapeutic endeavors and careful consideration in counseling patients and parents. Funding: Please list any funding that was received in support of this abstract.: No direct outside funding was received for this project.