• Auto-injection;
  • Benzodiazepines;
  • Clinical trial;
  • Cluster seizures;
  • Serial seizures



A diazepam auto-injector (AI) has been developed for intramuscular administration to treat acute repetitive seizures (ARS). The objective of this study was to evaluate the efficacy and safety of the diazepam AI when administered by caregivers to control an episode of ARS.


In this phase III, randomized, double-blind, parallel-group, placebo-controlled, multicenter study, subjects with epilepsy on a stable antiepileptic drug regimen who required intermittent medical intervention to control ARS were randomized 1:1 to the placebo AI or the diazepam AI group. Subjects were stratified according to age (2–5, 6–11, ≥12 years). Dose (5, 10, 15, or 20 mg) was based on age and weight. A single dose of study medication was dispensed to be administered by caregivers in an outpatient setting when required. The primary end point was time to next seizure or rescue from 15 min to 12 h postdose. Secondary end points included rescue medication use, number of seizures postdose, caregiver and physician treatment assessments, and safety measures.

Key Findings

Of 234 subjects randomized, 81/110 in the placebo AI group and 82/124 in the diazepam AI group were included in the intent-to-treat analysis. Baseline characteristics were similar for both groups. Time to next seizure or rescue was significantly longer in the diazepam AI group compared with the placebo AI group, with a hazard ratio of 0.55 (95% confidence interval (CI) 0.34–0.88; p = 0.012) for diazepam AI versus placebo AI, adjusted for age group. The 25th percentile for time to the next seizure or rescue was 1.18 h (95% CI 0.38–2.03) for placebo AI and 2.70 h (95% CI 0.48–11.42) for diazepam AI; the median was 5.9 h for placebo AI and was inestimable for diazepam AI due to the low number of events experienced by subjects in that group. The proportion of subjects using rescue medication postdose was 30% (24/81) placebo AI versus 17% (14/82) diazepam AI (p = 0.066). An event (seizure or rescue) occurred in 55.6% of subjects in the placebo AI group and 35.4% in the diazepam AI group. The number of seizures experienced during the 12-h postdose period was significantly lower for diazepam AI (median 0.0) compared with placebo AI (median 1.0; p = 0.010). Treatment-emergent adverse events (TEAEs) were reported in 44% (35/79) of subjects in the placebo AI group and 42% (34/81) in the diazepam AI group. The most common TEAEs reported were injection site pain (15% placebo AI, 17% diazepam AI) and injection site hemorrhage (6% placebo AI, 5% diazepam AI).


The diazepam AI was significantly more effective than placebo AI at delaying the next seizure or rescue. Secondary efficacy end points were generally supportive of the primary outcome. Diazepam AI administered by trained caregivers was effective for the treatment of ARS and was well-tolerated, with a safety profile similar to placebo.