Background: Epilepsy occurs in 70–90% of patients with tuberous sclerosis complex. We aimed to assess the efficacy and safety of adjunctive everolimus for treatment-refractory seizures associated with tuberous sclerosis complex in paediatric patients enrolled in the EXIST-3 trial, a double-blind, placebo-controlled, randomised, phase 3 study. Methods: This post-hoc analysis focused on paediatric patients (age <18 years) in the EXIST-3 trial, which consisted of baseline (8 weeks), core (18 weeks), and extension phases (≥48 weeks) and was done at 99 centres in 25 countries worldwide. Briefly, patients with tuberous sclerosis complex-associated treatment-refractory seizures, who were receiving a stable dose of one to three antiepileptic drugs, were randomly assigned (1:1:1) to receive placebo, low-exposure everolimus (3–7 ng/mL), or high-exposure everolimus (9–15 ng/mL). Following the core phase, patients could enter the extension phase to receive everolimus at a targeted exposure range of 3–15 ng/mL up to 48 weeks after the last patient had completed the core phase. Efficacy endpoints were response rate (≥50% of reduction from baseline in average weekly seizure frequency) and median percentage reduction in seizure frequency during the 12-week maintenance period of the core phase, and at 12-week intervals throughout the extension phase. This study is registered with ClinicalTrials.gov, number NCT01713946. Findings: Between July 3, 2013, and May 29, 2015, 299 paediatric patients enrolled in the trial. In the younger subgroup (<6 years; n=104), 34 received placebo, 33 low-exposure everolimus, and 37 high-exposure everolimus; in the older subgroup (≥6 years to <18 years; n=195), 62 received placebo, 63 low-exposure everolimus, and 70 high-exposure everolimus. At the end of the core phase, response rate was higher in the treatment groups than placebo in both the younger subgroup (17·6% [6·8–34·5] for placebo vs 30·3% [95% CI 15·6–48·7; p=0·2245] for low-exposure everolimus vs 59·5% [42·1–75·2; p=0·0003] for high-exposure everolimus) and the older subgroup (12·9% [5·7–23·9] vs 27·0% [16·6–39·7; p=0·0491] vs 30·0% [19·6–42·1; p=0·0179]), as were median reduction in seizure frequency (12·3% [95% CI −10·1 to 24·8] vs 29·3% [95% CI 13·4 to 46·3; p=0·0474] vs 54·7% [43·5 to 73·1; p<0·0001] in younger patients; 13·5% [−3·0 to 26·8] vs 31·0% [16·1 to 42·9; p=0·0128] vs 34·8% [26·7 to 41·3; p=0·0006] in older patients). The efficacy persisted, with sustained seizure reduction after 1 year of treatment across both paediatric subgroups (response rate 48·9% [95% CI 38·1–59·8] for the younger subgroup vs 47·2% [39·3–55·2] for the older subgroup; median percentage reduction in seizure frequency 48·4% [95% CI 34·3–73·6] vs 48·0% [38·2–57·5]). At the cutoff date for the extension phase, grade 3 or 4 adverse events were reported in 45 (45%) younger patients (commonly pneumonia [n=16]) and 74 (38%) older patients (commonly pneumonia [n=8] and stomatitis [n=6]). Two deaths (pneumonia, which was suspected to be treatment-related, and sudden unexplained death due to epilepsy) were reported. Interpretation: Adjunctive everolimus resulted in sustained reductions in seizure frequency after 1 year and was well tolerated in paediatric patients with treatment-refractory seizures associated with tuberous sclerosis complex. Funding: Novartis Pharmaceuticals Corporation.
