Evaluation of Time to Event endpoints in randomized controlled trials of multiple antiseizure medications for uncontrolled epilepsy

Epilepsy is the third most common cause of neurological disability and death globally, but a third of patients are failed by antiseizure medications (ASMs). Randomized placebo-Control Trials (RCTs) for epilepsy have been the mainstay of trials for regulatory approval of new ASMs for decades. In a placebo-controlled RCT, participants were randomly assigned to receive either placebo, which is an inactive substance, or active treatment, which aims to improve their seizure frequency. In traditionally designed RCTs, patients with high seizure burden may remain on placebo or ineffective treatment for at least 5 months, even if seizures are frequent. While death was rare within RCTs at 6 per 1000 participant-years, participants randomized to placebo had a 5.8-fold increased risk of death, therefore it is critical to reduce participants’ exposure to ineffective treatments. These risks of participating in RCTs also contributes to trial recruitment issues.
The fundamental clinical research question this work addresses is if a Time-to-Event design can shorten participants’ exposure to placebo or ineffective treatment while also achieving the dual goals of evaluating treatment efficacy and safety. In a Time-to-Event design, participants receive treatment until they reach specific efficacy or safety endpoints. This proposal is important and original because it directly addresses concerns that have been barriers to adoption of the most promising Time-to-Event design, known as time to pre-randomization seizure count (T-PSC). In T-PSC, participants receive treatment until their post-randomization seizure count exceeds the average monthly PSC experienced during baseline. The expected societal impact is to have sufficient ad hoc evidence that T-PSC is a safe and effective method that could be applied in prospective RCTs. The expected impact of this research on patient care and human health is to reduce trial participants risk of death, trial cost, and improve participant recruitment. The rationale of this project is that by making RCTs safer, cheaper, and more efficient, we enable further development of novel treatments for epilepsy to reduce the impact of seizures and adverse effects of treatment on quality of life and mortality through all causes including Sudden Unexpected Death in Epilepsy (SUDEP). Both prior to and after this work, RCTs maintain the highest ethical standards to protect participants from adverse effects based on the data available at the time. If this re-analysis demonstrates that efficacy can be measured equivalently and safety can be improved with the T-PSC design, then our results may provide new data that could change RCT protocols directly.
We will perform this analysis by re-calculating the effectiveness and safety outcomes of as many RCTs for uncontrolled epilepsy as are available with the T-PSC and other Time to Event designs. We can do this by re-producing each trial’s original analysis with truncating data, where data after the Time to Event endpoint is not included. By demonstrating the effectiveness of T-PSC on as many historical RCTs as possible, we will build strong evidence that T-PSC could be applied prospectively.