The management of generalized convulsive status epilepticus remains in perpetual status controversius. There is consensus that the front-line agent should be a benzodiazepine, but little agreement beyond that. The much-awaited ESETT trial is a multi-center RCT comparing three anti-epileptic agents for ongoing status epilepticus (levetiracetam, fosphenytoin, and valproate). This trial will be the highest-quality evidence regarding second-line antiepileptics.
(Although the ESETT trial does include children, please note that the PulmCrit blog pertains to adult critical care only.)
ESETT trial overview
This is a trial of adults and children with persistent generalized convulsive status epilepticus despite benzodiazepine therapy.1 Baseline characteristics are as follows:
Patients were randomized to receive one of three second-line agents. The primary endpoint was resolution of seizure and improving mental status one hour after starting the trial medication. There was absolutely no difference in this endpoint (achieved in ~46% of all patient groups):
The composite endpoint of seizure control plus improvement in mental status may be a bit subjective (depending on exactly how “mental status improvement” is defined). I would have preferred the non-composite endpoint of seizure control after one hour. The authors did include that endpoint in the supplemental data and it, too, is exactly the same across all groups (60%). So, no matter how you slice or dice the data, the three groups have similar outcomes.
The study isn’t powered to detect rare adverse events. More patients in the fosphenytoin group did develop hypotension, which is consistent with the drug’s known side-effect profile. However, these differences are nonsignificant.
This is an extremely well-designed trial with clear results, so the study largely speaks for itself. In the remainder of this post, I will speculate as to potential clinical implications from this study.
#1) fosphenytoin is dead for status epilepticus
Among the second-line agents for status epilepticus, fosphenytoin is probably the most dangerous. It may cause hypotension or bradycardia if it is administered too rapidly, or in patients with underlying cardiac comorbidities. This is a major problem, because the nuances of a patient’s cardiac history can easily be lost in the chaos of status epilepticus. Likewise, status epilepticus is a high-stress situation where medications can easily be given a bit too fast. My experience has been that in a medical center using fosphenytoin as a front-line antiepileptic agent, about 1 patient per year will end up in the ICU due to hypotension or bradycardia from fosphenytoin. After switching to levetiracetam as the go-to antiepileptic, this problem disappears.
The traditional justification for using fosphenytoin in status epilepticus was that it was supported by stronger evidence than levetiracetam. Well, that’s no longer true. Given that fosphenytoin, levetiracetam, or valproate all have the same efficacy, it makes sense to pick the drug which is safest and easiest to give rapidly. And that drug is levetiracetam:
The other main reason to get rid of fosphenytoin is that achieving a therapeutic level can be challenging. Phenytoin interacts with dozens of medications. Furthermore, the drug is bound to albumin, so to really understand the drug level in critically ill patients you want a free fosphenytoin level (good luck getting that STAT at 2 AM). The pharmacologic quirkiness of phenytoin not only soaks up resources (e.g. lab draws, re-dosing, calls with pharmacy), it also introduces the risk of phenytoin toxicity.
#2) big levetiracetam doses will become standard
The authors went big with their levetiracetam dose (60 mg/kg, up to a max dose of 4.5 grams, infused over 10 minutes). This seems like a wise choice. Levetiracetam is extremely well tolerated, even at very high doses. Given that status epilepticus is a life-threatening process, it makes sense to dose levetiracetam like you mean it.
Interestingly, the most recent status epilepticus guidelines by the American Epilepsy Society have already adopted this exact same levetiracetam dose.2 The writers of this guideline most likely realized that after ESETT was published, its dose of levetiracetam would become standard. Importantly, though, the authors of the guideline agreed with ESETT on the use of this dose.
#3) all conventional anti-epileptic agents fail as 2nd line agents to break convulsive status epilepticus
Here is where things will get really controversial. Let’s start with a fundamental question. How long is it safe to leave an adult patient in generalized convulsive status epilepticus?
There are numerous risks involved in convulsive status epilepticus, for example:
- Permanent neurologic damage may occur at some point (nobody knows when, but it might be as early as 30 minutes).3 This may depend on the underlying brain disorder, so it’s probably impossible to establish any universal rule.
- Aspiration becomes increasingly likely over time.
- Rhabdomyolysis, hyperthermia, hyperkalemia, and acidosis gradually develop.
- Substantial cardiovascular stress occurs, which may lead to malignant arrhythmias and sudden death (e.g., three patients in ESETT suffered from cardiac arrest).
- Convulsive status epilepticus prevents other interventions from being performed (e.g. CT scans, lumbar puncture, trauma evaluation).
- The longer the seizures persist, the harder they become to break. Thus, a sluggish intervention strategy may cause more patients to develop super-refractory status epilepticus (which, unlike most cases of status epilepticus, can become a true management nightmare).
My opinion is that we should aim to always break the seizure in <30 minutes (discussed previously in 2014 and 2017). This is somewhat controversial, but I think most adult resuscitationists would agree that allowing a patient to be in ongoing generalized status epilepticus for >45 minutes is suboptimal and >60 minutes is definitively scary. Folks on twitter feel likewise:
By any of these standards, conventional anti-epileptic agents failed to gain seizure control within a safe time interval. ~40% of patients in all groups were continuing to seize an hour after administration of the study drug. The median duration of seizure at enrollment was about an hour, so this 40% of patients were seizing on average for over two hours. Yikes.
I don’t believe that this failure represents a failure of the investigators. On the contrary, the investigators and centers involved in ESETT are clearly outstanding. Furthermore, care received within an RCT is often superior to routine care (due to greater focus and energy dedicated to it). The issue here is that the traditional paradigm of sequential administration of different anti-epileptic drugs is simply too slow to achieve rapid seizure termination. For example, in an idealized fantasyland where zero delays to care exist, this strategy would still fail to control seizures in thirty minutes:
This timeline shows what would happen if every drug were given as early as physically feasible. In this scenario, only 60% of patients would respond to levetiracetam, so 40% of patients would still be seizing after 30 minutes. So even this fantasyland scenario fails to guarantee seizure control in <30 minutes.
There is only one way to potentially achieve seizure control in <30 minutes using conventional first- and second-line agents. Once patients have been seizing continuously for >5 minutes, then they have status epilepticus and they will require a maintenance anti-epileptic agent (i.e., even if the lorazepam breaks the seizure, the patient will still need levetiracetam as a maintenance anti-epileptic). Therefore, there’s no benefit to delaying administration of the levetiracetam for a patient in status epilepticus. Thus, one could imagine the following accelerated sequence of events:
This compressed sequence of events could achieve seizure control in <30 minutes, while avoiding intubation whenever possible. This would require immediate administration of levetiracetam doses (pragmatically speaking, the drug would need to be kept ready in the emergency department and/or ICU). This degree of speed is unlikely to be achievable in most contexts.
Realistically, there are two ways I can imagine controlling status epilepticus reliably in <30 minutes. The first strategy is to use propofol plus ketamine as a second-line antiepileptic agent (along with intubation, as shown below). This is generally my go-to strategy for status epilepticus, which seems to be safe and effective:
An early-intubation strategy will result in intubating some patients who might not have required intubation. That’s acceptable. Such patients are generally easy to extubate the following morning. Early intubation may theoretically avoid several complications (e.g. aspiration, cardiac arrest, brain damage, super-refractory status epilepticus), all of which are much harder to manage.
The only other way to guarantee seizure control in <30 minutes could be to attempt to give ketamine as a second-line antiepileptic, without intubation (see below). Ketamine is an excellent anti-epileptic agent, and it shouldn’t suppress respiration. Thus, ketamine could be used to break status epilepticus without requiring intubation. There is some evidence to support this strategy from the pediatric and basic science literature.4,5 Currently, I can’t recommend this strategy (due to lack of adequate evidence), but it is intriguing. Further research on second-line agents in status epilepticus should hopefully include ketamine.
- ESETT is a landmark trial which represents the highest quality evidence available for comparing second-line antiepileptic agents in generalized convulsive status epilepticus.
- All three traditional agents had exactly the same efficacy (fosphenytoin, levetiracetam, and valproate). It therefore makes sense to choose the agent which is safest and easiest to give (e.g. with the fewest contraindications, since we frequently must treat patients without knowing their full medical histories). This agent is arguably levetiracetam (and definitely not fosphenytoin).
- Large doses of levetiracetam were used in this study (60 mg/kg, up to a maximal dose of 4.5 grams IV). This same exact dose was recommended by the most recent guideline on status epilepticus. If you use this dose, you’re in good company and have adequate literature to back you up.
- All conventional anti-epileptic agents failed to reliably achieve rapid seizure control (with 40% of patients seizing for over an hour after administration of the study medication). Achieving rapid seizure control probably requires a radically different strategy (with either ketamine, or ketamine plus propofol, as the second-line agent after benzodiazepines).
- (Please note that the PulmCrit blog applies only to adult critical care.)
related
- PulmCrit: Resuscitationist's guide to status epilepticus (2017), Rapid sequence termination of status epilepticus (2014)
- Status Epilepticus with Tom Bleck (EMCrit 155)
- SMACC Chicago talk by Tom Bleck
- Status Epilepticus: Emergency management & rationale (First10EM, Justin Morganstern)
- Status Epilepticus Critical Care Compendium (LITFL)
references
- 1.Kapur J, Elm J, Chamberlain J, et al. Randomized Trial of Three Anticonvulsant Medications for Status Epilepticus. N Engl J Med. 2019;381(22):2103-2113. doi:10.1056/NEJMoa1905795
- 2.Glauser T, Shinnar S, Gloss D, et al. Evidence-Based Guideline: Treatment of Convulsive Status Epilepticus in Children and Adults: Report of the Guideline Committee of the American Epilepsy Society. Epilepsy Curr. 2016;16(1):48-61. doi:10.5698/1535-7597-16.1.48
- 3.Zaccara G, Giannasi G, Oggioni R, et al. Challenges in the treatment of convulsive status epilepticus. Seizure. 2017;47:17-24. doi:10.1016/j.seizure.2017.02.015
- 4.Ilvento L, Rosati A, Marini C, L’Erario M, Mirabile L, Guerrini R. Ketamine in refractory convulsive status epilepticus in children avoids endotracheal intubation. Epilepsy Behav. 2015;49:343-346. doi:10.1016/j.yebeh.2015.06.019
- 5.Niquet J, Baldwin R, Norman K, Suchomelova L, Lumley L, Wasterlain C. Midazolam-ketamine dual therapy stops cholinergic status epilepticus and reduces Morris water maze deficits. Epilepsia. 2016;57(9):1406-1415. doi:10.1111/epi.13480
- Pulmcrit wee: The cutoff razor - April 15, 2024
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
Josh, As always, thanks for the great analysis. I definitely read your multiple disclaimers about being an adult focused blog, but I want to give my pediatric perspective. On the subject of levetiracetam availability – In our ER, we keep the vials in our Omnicells. The medication is prepped bedside to a final concentration of 50 mg/mL then pushed over 5-ish minutes based on the study by Wheless et al. This practice reduces time to administration and total administration time for levetiracetam which I think makes your simultaneous administration option more feasible. Here is why I worry about early intubation… Read more »
thanks! a couple thoughts, 1) I’m intensely jealous of having vials of levetiracetam stocked at the site of care (ED, ICU, helicopter, etc). Hopefully this will become standard practice in the United States as well. Keeping levetiracetam in pharmacy and trying to get it to the unit in time to treat status epilepticus just doesn’t work logistically. I think stocking levetiracetam vials at the site of care and giving it simultaneously with BZD as a front-line could avoid some intubations. 2) Yeah, kids might be different (e.g. they don’t have coronary disease or risk of malignant arrhythmia). I really have… Read more »
EcLiPSE (PMID 31005385) and ConSEPT (PMID 31005386) were two RCTs just published in the Lancet a few months ago, and looking at a pediatric-only population, levetiracetam and phenytoin had similar outcomes too. However, they used 40 mg/kg levetiracetam over 5 min with max 2500mg-3000mg – due to PMID 16886975 is my guess.
I’m a little confused that only ~20% of patients were intubated after the initiation of the second line agent but 40% were still seizing at this time point? They’re watching a patient seize for an hour after the second line agent without intubating them?
I think so. Opinions on management of status epilepticus differ, and some folks will allow patients to seize for long periods of time. I think ED & ICU folks are much faster to intubate these patients because they’re more comfortable with intubation, whereas neurologists tend to view intubation as a failure.
I am concerned about the medico-legal concerns regarding following benzos up with prop/ketamine/ as second-line agents for seizure abortion, All it takes is one bad airway in a young person with anoxic injury and the ultimate peer-review/medico-legal question is why didn’t you use a normal second line AED medication before going to intubation for seizure control. That would be a difficult one, I think. These seizure intubations are often chaotic with high risk of aspiration. There are obviously aspiration risks while the patient is actively seizing as well, but through the legal eyes, I see that practice as problematic. Intubating… Read more »
I don’t think the medicolegal standard of care is that we must watch a patient aspirate with generalized seizure until giving a certain number of drugs before we intubate. The foundation of EM and CCM is that if we think the patient isn’t protecting the airway we should intervene. I don’t think this would be difficult to defend in court. Honestly I think we might be more likely to get into legal trouble by acting in a delayed fashion (failure-to-act). But the best thing is really just to provide the best possible care that you can and lets the cards… Read more »
Thanks for the excellent write up. Appreciating that the trial isn’t powered for the adverse events, shouldn’t we be a little cautious of the fact that at this hitherto relatively untested dose of 60mg/kg, 5% of the keppra group died. Co-incidence – probably but is there confidence this is a safe dosage?
I will address this in an upcoming blog on mortality endpoints and power (which in some ways will be a continuation of a prior blog on the same topic https://emcrit.org/pulmcrit/mortality/). When you calculate the numbers on these comparisons, they are nowhere close to statistical significance. The mortality differences are almost certainly a fluke and don’t reach a level of evidence that should affect clinical practice.