Real-World Analysis of Antiepileptic Drugs

Trevor Resnick, MD: This is an interesting way of looking at effectiveness through retrospective claims analysis, because it’s prone to a lot of variables that may affect outcomes. Nevertheless, it is a mechanism that actually the insurance industry does look at a lot, so it’s important for us to be aware of the data. And in this study, what [the investigators] did is they looked at the claims analysis comparing hospitalization rates for perampanel and lacosamide. Now, the reason for them looking at both is [that] anytime you try something new, it causes what is called active management. So you’re more involved in the patient, you’re more involved discussing [adverse] effects [and] effectiveness. And just that active management sometimes affects outcome. And this way, by comparing perampanel [with] another effective antiepileptic drug, such as lacosamide, it really creates a level playing field that takes out the bias of active management. And taking into account the variables that we did discuss, for both lacosamide and perampanel, there was a meaningful reduction in epilepsy-related hospitalization rates following the initiation of either perampanel [or] lacosamide, with perampanel being more effective than lacosamide,…a gratifying result for both medications.

R. Edward Faught, MD: There are data that can’t be obtained from a typical clinical trial involving a few hundred patients that you may be able to obtain from large national databases. And we’re particularly interested in patterns of any epileptic drug use and how they work out, in terms of real-world use of these drugs. So we use something called a Symphony database, which encompasses about 280 million lives per year in the United States, and it’s a claims database. The data [are] put in by hospitals, by physicians, by pharmacies, and so forth.

So we use that to look at the use of 2 relatively new antiepileptic drugs: lacosamide and perampanel. And since you can’t get data on actual seizure occurrence from these databases, we use a proxy, which is hospitalizations, with the idea that people who had more hospitalizations probably had poor seizure control. That’s something that’s commonly used in this type of study.

So we looked at all patients who were prescribed 1 or the other of these drugs over a 2-year period, from 2014 to 2016, and then matched groups between the 2 drugs on comorbidities, age, other things like that. And then we looked at how frequently they were admitted to the hospital over the next year. and that gave us a proxy for comparing the effect of the drugs on hospitalization rates and presumably control of epilepsy.

Well, the results were that starting a new drug is a good thing to do. If people are having uncontrolled seizures starting, doing something rather than nothing is always a good idea. And both drugs reduced the propensity for hospitalization over the next year. This is 1 of the only studies to compare 2 drugs head-to-head in this kind of a situation. And we found that the perampanel addition did a little better than the lacosamide addition. The difference was not huge. It was for people who’d had a previous hospitalization in the year prior. These presumably were the sicker patients. The perampanel reduced the chance of another hospitalization by about half, and the lacosamide, by about 40%. So there’s about a 10% difference, but it was statistically significant.