In 2014 I wrote a post suggesting an aggressive, streamlined approach to status epilepticus involving early intubation. The fundamentals of that post remain valid. However, much has changed over the last few years. This post aims to refresh and extend the prior post. It will also serve as a reference to explain my algorithm for status epilepticus in the upcoming Resuscitationist's Crisis Manual.
Overview & philosophy
There are many ways to manage status epilepticus. Rather than focusing on what may be the theoretically ideal, this post explores an approach which is effective, safe, and feasible. It's easy to write an algorithm that looks pretty on paper, but harder to create one that works in a crisis. For example, it's easy to write “administer fosphenytoin after 10 minutes.” However, the logistics of ordering the drug, receiving it from pharmacy, infusing it at the correct rate, and allowing the body to metabolize it into active phenytoin can take an hour.
The rationale for aggressive treatment of generalized status epilepticus was described previously here. In short, best outcomes rely upon rapid seizure control. The longer the seizure continues, the more refractory it becomes to therapy. The duration of status epilepticus which may cause permanent brain damage is unknown, with experts currently suggesting thirty minutes (Zaccara 2017). Aside from the brain, persistent status epilepticus may cause aspiration, hyperkalemia, rhabdomyolysis, hyperthermia, myocardial infarction, and arrhythmia.
The role of intubation is controversial. Neurologists tend to view intubation as a form of “treatment failure,” reflecting inability to treat the seizure with traditional anti-epileptic agents. My opinion is that intubation is often a key therapeutic intervention to control the seizure and prevent complications. Patients who are intubated early usually have good neurologic outcomes allowing extubation a day or two later. Alternatively, patients who are intubated after long periods of uncontrolled seizures are at increased risk for super-refractory status epilepticus, poor neurologic outcomes, aspiration pneumonia, and longer time on the ventilator.
The backbone of this algorithm is constructed from medications that are immediately available in any critical care arena (lorazepam, propofol, and ketamine). If levetiracetam arrives from the pharmacy that's a bonus, because the algorithm will work just fine without it. Utilizing immediately available medications which can be bolused is the only way to reliably achieve rapid seizure control.
Defining generalized status epilepticus
This post is about generalized convulsive status epilepticus, which causes loss of consciousness and diffuse muscular activity. This must be distinguished from continuous partial status epilepticus, a rare condition involving ongoing partial seizures (e.g. movement of an arm, without loss of consciousness). Partial status epilepticus may be managed in a less aggressive fashion.
Generalized status epilepticus is currently defined as either:
- Ongoing convulsive seizure > 5 minutes
- Recurrent seizures without normalization of consciousness between seizures.
The algorithm described here is designed for patients with persistent, ongoing generalized seizure activity. For a patient with recurrent seizures who isn't actively seizing, a less aggressive strategy may be used.
Benzodiazepine: First-line therapy
For patients with IV access, lorazepam is generally accepted as the first-line therapy. This is based largely on the landmark VA Cooperative Trial, a prospective RCT comparing lorazepam to other antiepileptics including phenobarbital and phenytoin.
The dose of lorazepam used in the VA Cooperative Trial and most RCTs is 0.1 mg/kg (Trinka 2015)(1). This is arguably the most evidence-based dose. However, guidelines usually recommend a dose of 4 mg IV, with a repeat dose if needed.
The algorithm above utilizes 0.1 mg/kg of lorazepam, which may seem like a lot. However, this dose reduces the rate of respiratory depression experienced during status epilepticus compared to placebo, indicating that benzodiazepine is less dangerous than ongoing seizure activity (Glauser 2016). It should also be noted that pediatric studies and guidelines recommend a dose of 0.1 mg/kg for children. It makes little sense to use the same dose for a 40-kg child and all adults:
For patients without intravenous access, evidence supports a dose of 10 mg intramuscular midazolam. The RAMPART study found that up-front intramuscular midazolam was more effective than first attempting to get IV access and then giving IV lorazepam.
Causes of seizure that require immediate treatment
There are dozens of causes of status epilepticus. However, for immediate management two are the most important: hypoglycemia and hyponatremia.
Hypoglycemia must be excluded in any patient with seizure or altered mental status. This can generally be accomplished by measuring a fingerstick glucose. If fingerstick glucose can't be obtained or is borderline, just give IV dextrose (e.g. 1-2 ampules of D50W).
Hyponatremia may also require immediate treatment. If electrolytes aren't known, they should be measured (ideally with a point-of-care device which will provide rapid results). Hyponatremia can be treated by bolusing ~150 ml of 3% saline, with a repeat bolus for persistent seizures. However, this often takes a while to obtain from pharmacy. Two ampules of hypertonic bicarbonate provide a similar amount of hypertonic therapy, with the advantage that they are immediately available.
Intubation using propofol
Propofol as a second-line antiepileptic agent
The rationale for propofol as a second-line anti-epileptic (rather than waiting for a conventional anti-epileptic agent) was explored in detail in a prior post. In short, most patients who fail benzodiazepine will require intubation eventually (figure below). Delaying intubation to allow for a trial of anti-epileptic risks prolonging the seizure and increasing associated complications.
Propofol is a potent anti-epileptic agent. When bolused during rapid sequence intubation, this will usually break the seizure. Propofol must subsequently be infused at a moderate dose (e.g. 50-80 mcg/kg/min) to maintain control of the seizure.
The advantage of propofol is that it is rapidly titratable. Thus, a high dose of propofol may be used initially to gain control of the seizure. Once the seizure is controlled and the dust has settled, propofol may be down-titrated as needed (2).
The main disadvantage of propofol is that it causes hypotension. This is generally manageable (e.g. with a low-dose norepinephrine or phenylephrine infusion). However, for a patient with severe shock, propofol may not be safe. These patients may be managed using midazolam instead (with a loading dose of 0.2 mg/kg, followed by an infusion of 0.1 mg/kg/hr)(3).
Prolonged infusion of propofol at high rates can cause propofol infusion syndrome (a highly morbid condition involving bradycardia, lactic acidosis, and shock). This may be avoided by using propofol infusion rates below 83 mcg/kg/min (<5 mg/kg/hr) and serial monitoring of triglyceride levels (4).
Ketamine as an adjunctive antiepileptic agent
Ketamine is a powerful anti-epileptic agent. For example, ketamine has shown efficacy in status epilepticus refractory to a variety of other medications. Consequently, some guidelines have added ketamine as a possible treatment for super-refractory status epilepticus (Fung 2017).
The combination of ketamine and propofol should theoretically provide synergistic anti-epileptic activity. Propofol stimulates the GABA receptors (the main inhibitory neurotransmitter in the brain), whereas ketamine inhibits NMDA receptors (a major excitatory neurotransmitter). The combination of these two effects causes a profound decrease in CNS activity. Ketamine combined with propofol (“ketofol”) was effective in published series of patients with super-refractory status epilepticus (Sabharwal 2015, Hofler 2016)(5).
Choice of paralytic for intubation
The ideal paralytic for status epilepticus is controversial. Rocuronium is an excellent paralytic for intubation, but it will obscure the neurologic examination for about an hour. Prolonged paralysis could create a situation where the patient is having persistent electrical seizures (causing brain damage), without observable movements. There are numerous reasonable approaches to this, including:
- Intubation with succinylcholine (if there is no contraindication to this). Caution is required, however, because ongoing status epilepticus may cause hyperkalemia after >20-30 minutes.
- Intubation with rocuronium, followed by reversal of rocuronium with sugammadex to obtain a neurologic examination.
- If you don't have sugammadex, intubation with rocuronium may still be used. Following intubation, consider giving high-dose propofol infusion, additional ketamine, and a conventional anti-epileptic agent to provide extra protection against recurrent seizure while the patient is paralyzed.
- Intubation without a paralytic (a bolus of 1.5-2 mg/kg propofol will generally produce good intubating conditions, albeit for a short time).
Conventional anti-epileptic drug
Equipoise
According to the most recent guidelines, first-line conventional antiepileptic drugs include fosphenytoin, valproic acid, or levetiracetam (Glauser 2016). Among these, there is no clear evidence regarding which is the most effective. The ESETT trial is currently recruiting patients to clarify this.
Reasons not to use fosphenytoin
phenytoin can no longer be considered as the first choice. While phenytoin does not seem more effective, and perhaps is less effective than valproate, it is certainly associated with several risks and is not easy to administer. –Zaccara 2017
Fosphenytoin has traditionally been regarded as the first-line antiepileptic agent. This isn't because of evidence of superiority, but rather because there is the greatest amount of prior experience with its use (status quo bias). Given problems with fosphenytoin outlined below, many institutions are moving away from it.
- Fosphenytoin has more adverse drug effects than other agents, especially hypotension and bradycardia (I've encountered two patients who have experienced bradycardiac arrest following administration of fosphenytoin).
- Fosphenytoin has sodium-channel blocking effects, which could be problematic in the setting of status epilepticus due to sodium-channel blocker intoxication (e.g. tricyclic antidepressant). However, it is often impossible to know immediately if a patient's seizure has a toxicological etiology.
- Fosphenytoin has numerous drug-drug interactions, necessitating ongoing measurement of drug levels and perpetual dose titration. Spending five minutes on rounds deciding on the phenytoin dose is a distraction from more important issues (6).
- Even patients who respond well to phenytoin will usually get transitioned to levetiracetam eventually. Instead of exposing the patient to two different drugs, why not start with levetiracetam in the first place?
- After infusion fosphenytoin must be converted by the body into active drug (phenytoin) before it can work. This is a pharmacokinetic drawback compared to valproic acid and levetiracetam, which likely achieve therapeutic levels in the brain faster.
Levetiracetam vs. valproic acid?
Both levetiracetam and valproic acid are excellent options. Levetiracetam is often favored because:
- Levetiracetam has nearly no contraindications, which makes it a good choice when giving a medication emergently to someone you know little about. In contrast, valproic acid is contraindicated in some situations (hepatic disease, urea cycle disorders, mitochondrial diseases).
- Levetiracetam is easy to administer (doesn't require monitoring of drug levels).
Timing of conventional anti-epileptic drug
Even if the seizure is controlled with lorazepam, the patient should still receive a conventional anti-epileptic drug. Lorazepam will only provide protection against seizures for a few hours. Failing to provide ongoing anti-epileptic therapy leaves the patient at risk for recurrent seizures.
Therefore, any patient who has been seizing for >5 minutes should receive a conventional anti-epileptic drug. There is no merit to delaying this. Ideally, the conventional anti-epileptic agent would be given at the 5-minute mark (simultaneously with the lorazepam).
In reality, ordering the anti-epileptic drug from pharmacy, receiving it, and infusing it will often take ~20-40 minutes. Using the above algorithm, the levetiracetam will usually arrive from pharmacy after the seizure has already been controlled with propofol. Thus, in practice the role of the levetiracetam is generally to prevent recurrent seizure.
Wrapping it up
Following seizure control, additional investigation may be required to determine the cause of the seizure (e.g. CT scan, MRI, LP, toxicology labs, anti-epileptic drug levels, etc.). Any causes and contributory factors should be treated.
Video EEG is helpful to monitor therapy. It is unknown whether it is sufficient to simply control seizures, or whether deeper levels of anesthesia might be preferable (e.g. inducing a burst-suppression or completely flat EEG pattern). The optimal duration of anesthesia is also unknown. A reasonable approach may be to start by suppressing seizure activity for a day, and then trying to wean off sedation (7). With this strategy, most patients will only require a day of mechanical ventilation (8). Use of high-dose maintence antiepileptics (e.g. levetiracetam) throughout this period may avoid recurrent seizures.
Comparison to published guidelines
The most recent guidelines by the American Epilepsy Society are shown here (Glauser 2016):
These guidelines note that “depending on the etiology or severity of the seizure, patients may go through the phases faster or even skip the second phase and move rapidly to the third phase, especially in sick or intensive care unit patients.” Therefore, the algorithm proposed above is fundamentally consistent with these guidelines.
Perhaps the largest difference between the algorithm above and these guidelines is the adjunctive use of ketamine. When I suggested using propofol plus ketamine for anesthetic induction in 2014, this was a bit more radical. Since then, other authors have also suggested that ketamine should be used earlier in the course of status epilepticus (Zeiler 2015). Given its outstanding safety profile, there is no reason to reserve ketamine as a last-ditch intervention for super-refractory status epilepticus. For example, Ilvento 2015 reported on a series of children where ketamine was used to control status epilepticus without requiring intubation (9).
- Ongoing generalized convulsive seizures are an immediate life-threat that may cause brain damage, aspiration, rhabdomyolysis, hyperkalemia, arrhythmia, and hyperthermia.
- A streamlined algorithm is proposed to achieve rapid seizure control using immediately available medications.
- Patients who fail to respond to benzodiazepine will frequently require intubation and sedative infusion to control their seizures. Early intubation in these patients may expedite seizure control and avoid complications.
- Propofol and ketamine are powerful anti-epileptic agents. Emerging evidence suggests that they work synergistically by affecting GABA and NMDA receptors, respectively.
- The best conventional anti-epileptic agent is unknown and currently under investigation in the ESSET trial. Fosphenytoin has the most side-effects, so for now levetiracetam or valproic acid seem like the best choices.
Related
- FOAM
- Rapid Sequence Termination of Status Epilepticus (PulmCrit)
- Status Epilepticus with Tom Bleck (EMCrit 155)
- SMACC Chicago talk by Tom Bleck
- Status Epilepticus (First 10 in EM)
- Status Epilepticus Critical Care Compendium (LITFL)
- Seizure answers by EM Lyceum
- Seizures (EM Basic)
- Current Consensus Guideline
- Glauser T et al. Evidence-based guideline: Treatment of convulsive status epilepticus in children and adults: Report of the guideline committee of the American Epilepsy Society. 2016 16: 48-61. (Open access)
Notes
- The VA Cooperative Trial provided lorazepam as an infusion at up to 2 mg/min. Since lorazepam typically has an onset of action of 5-10 minutes, providing it as an infusion over 4-5 minutes probably doesn't add much benefit compared to providing the entire dose as a bolus (the drug won't have significant effect until after the infusion is over anyway). This is consistent with the results observed in the VA Cooperative study, specifically nearly all subjects in the lorazepam arm ended up receiving the full 0.1 mg/kg dose (suggesting that neither clinical benefit nor harm were observed while the lorazepam was infused).
- There is no clear evidence comparing different anti-epileptic maintenance infusions to one another. One retrospective study of 20 patients found a non-significant correlation between mortality and use of propofol (Prasad 2001). It should be noted that patients in this study received up to 24 mg/kg/hr of propofol, a very high dose which might have increased the risk of complications. This is a small retrospective study which does not prove that moderate doses of propofol are dangerous.
- For refractory seizures, midazolam may be re-loaded (up to a total loading dose of 2 mg/kg) and the infusion increased (as high as 1 mg/kg/hr). However, using these doses of midazolam may make it very difficult to wake the patient up. Use of such high doses of midazolam should probably be performed in consultation with neurology.
- Serial monitoring of triglyceride levels may help prevent propofol infusion syndrome, by detecting early metabolic derangements which signal deterioration if the propofol isn't stopped. Ideally the ICU should have a protocol for using propofol which incorporates these measures automatically.
- Ketamine should in theory should also work synergistically with other GABA-agonists (e.g. benzodiazepines and barbiturates).
- For more on this, see decision fatigue as discussed here by Scott Weingart.
- If patients seize while sedation is being weaned, then they need re-initiation of deep sedation as well as addition of another anti-epileptic agent (e.g. lacosamide).
- This is going to depend on the patient population and other active medical and neurologic problems that these patients may have. For example, a patient with seizure due to severe meningitis or encephalitis will often need more than a day on the ventilator. However, patients without other active problems (e.g. status epilepticus due to antiepileptic nonadherence) who are treated aggressively with rapid lysis of the seizure can generally be extubated rapidly.
- This is a brilliant concept, and hopefully something that will eventually gain traction with adults as well. Currently there isn't enough data to support widespread use of this strategy in adults.
- 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
Two comments:
1) LevETIRAcetam is misspelled throughout the article, you are missing the “e” 🙂
2) LevETIRAcetam should be in all automated dispensing cabinets, it has been available as a premixed infusion for a while now, thus preventing the delay from the pharmacy as we can take the pharmacy out of the equation. I think that the biggest hurdle would be to make sure that not everyone gets 1000 mg but gets a weight based dose!
Thanks for this great review!
Excellent points. Damn drug companies intentionally make the generic names difficult to spell, in order to encourage us to use the tradenames. Yes, that’s right, I’m blaming my inability to spell on big pharma.
In Australia we use Midazolam Jody ICU nurse
Midazolam is a fine choice. There is no high-quality evidence regarding the selection of propofol vs. midazolam. So you’re on solid ground using either. Personally I prefer propofol due to its pharmacokinetics. Propofol is nice for anesthetic induction because it has extremely fast onset which will typically lyse the seizure immediately. The fast-onset is also helpful because if propofol *fails* control the seizure promptly then you know that you need a different or additional agent (rather than re-loading midazolam over and over again, which can take a while before realizing that it’s not going to work). Finally, propofol is much… Read more »
A standard bottle of sodium bicarbonate in Australia tends to be 100ml of 8.4% (100 mmol). One bottle of this rather than two would be roughly the same sodium load as 150 ml of 3%NaCl.
that’s not an amp of bicarb… THIS is an amp of bicarb!
(sorry, Crocodile Dundee impression hard to do on blog)
Could you comment some more on the dosing of Keppra? The AES guideline for 60mg/kg at Level U evidence piqued my interested. It’s not something I’ve ever done in the ER. Is that dose commonly used in the neuro ICUs for other reasons?
There is good evidence that big doses of keppra (e.g. 3-4 grams iv) are safe. I’m not aware of any specific evidence that higher doses are more effective than intermediate doses (e.g. 3 grams vs. 4 grams). However, given how dangerous status epilepticus is, it seems reasonable to use a big dose. There is potentially a pharmacokinetic advantage to using a big dose (will achieve therapeutic levels in the brain faster). There isn’t any solid evidence for this dose currently. However, this is the dose being used in the ESETT trial. Following the publication of ESSETT, this dose will probably… Read more »
Interested to hear why you chose to put BZD treatment before putting r/o hypoglycemia. Makes most sense to me to r/o hypoglycemia and treat it if it’s present, should take precedence over administering BZD.
Although steps #2-5 are ordered sequentially, in reality these will often occur near-simultaneously. There are usually enough team members available to distribute these tasks to different people. Hypoglycemia is a rare cause of seizure (3% in one series, Ann Emerg Med 19:373). So if you don’t have enough staff to perform these tasks simultaneously, it might make sense to focus on the benzodiazepine (which will benefit ~100% of patients) and delay the finger-stick a minute (since it will benefit only ~3% of patients). Ultimately, either order is OK as long as you attend to both tasks promptly. If you are… Read more »
Hi Josh! Similar to another reader, I’m interested in the dosing of levetiracetam suggested in your post (and indeed, the SE guidelines). I believe this data is primarily derived from pediatric patients, and anything I’ve read for adults seems to suggest that doses above 3000 mg/day don’t confer any additional benefit (http://www.sciencedirect.com/science/article/pii/S1525505006000084). I think it’s a bit excessive to be recommending this dose broadly in adults, and I was discouraged to see it in the SE guidelines. For now, I’ll stick with recommending 20 mg/kg as a loading dose in adults.
I like big keppra doses, I cannot lie. Even for me, this is a bigger dose than I would have the courage to give on my own volition (without these guidelines). However, this is the dose used in the ESETT trial and recommended in the 2016 American Epilepsy Society guidelines. The people behind these efforts include the best epilleptologists in the world (Bleck, Lowenstein, Treiman, etc). So I think anyone using this dose is on reasonably solid ground, literature-wise. I’m happy to go against guidelines when there is good evidence to support doing so. However, in an evidence-free hazy zone,… Read more »
Also don’t forget that status may be secondary to eclampsia which can present up to 3 weeks post partum. Consider magnesium in females of child bearing age.
Completely agree.
Doctor…Is there any reason to not put in an IO rather than an IV in an actively seizing patient? It can be done much faster and easier and allows for use of lorazepam, and D50 should that be warranted.
these patients definitely need some sort of access (IO or IV). However the RAMPART study showed that giving IM midazolam without any delay is superior to getting IV access and then giving IV lorazepam, so this is probably the best evidence-supported therapy currently. Ideally midazolam would be given intramuscularly simultaneously with efforts to obtain access.
What is IV/IO access is established? Any difference in efficacy of Midazolam IV vs Ativan IV?
Thanks for a great post!
Great post and I look forward to the weekly updates as it helps to guide my own critical care practice. My question for you: Is this algorithm applicable to patients with non-convulsive status epilepticus as well? Thanks for your time.
Treatment of NCSE is generally similar to conclusive status epilepticus, so I think you could use the same medications. The diagnosis of NCSE can be harder to make, though, so it might be a good idea to get a neurologist’s impression and/or EEG confirmation of NCSE before pulling the trigger on management.
How do you see the role of thiopentone for induction?
Unfortunately I don’t have thiopental in the US. It sounds like it would be a good choice, fairly similar to propofol (powerful anti-epileptic agent with tendency to drop the blood pressure).
Great post. Do you know much about nonconvulsive status epilepticus, and specifically how the treatment choice and AED selection might be the same as or different from that of convulsive status? Have you ever read anything about ketamine in NCSE? I can’t find much. Thanks!
Why not levetiracetam plus valproic acid
why not midazolame as first line benzodiazepine
Good morning, friends. Is Dr. Farkas aware of the recent (2016) RCT by Navarro et al on levetiracetam in status epilepticus? Citation follows. This double-blinded RCT, the only one of its kind, was published after the most recent guidelines had been finalized, and it has persuaded me that levetiracetam is of no value as an adjunct to benzodiazepines in the management of status epilepticus. Instead, this trial suggests levetiracetam should be only ordered if the status has conclusively resolved before the clinician has had the chance to order the conventional AED. The regimen used in the trial was 2500 mg… Read more »
I agree with you that SAMUKeppra is a bit worrisome, and it would have been nicer to see keppra do better. However, numerous methodologies limitations prevent me from changing practice on the basis of this study. SAMUKeppra was stopped early based on interem analysis, leaving it underpowered. Stopping studies early due to futility or efficacy can lead to aberrant results and insufficient power. for example, on Table 2 the primary outcome (seizure control) was achieved in 50/68 patients in the keppra group vs. 57/68 patients in the control group. So that’s a 10% reduction in seizure control in the Keppra… Read more »
Thank you for your reply. I agree that your criticisms are valid.
Dear Josh, what a fantastic overview you gave us here. This approach makes much sense. Can we pick your brain on your choice of second line anti-epileptic drugs when a patient is already on a chronic dose of valproic acid or LevETIRAcetam. Would this fact influence your choice?
Great question. I’m not sure but here are some general concepts to work with. First thing here is that if they’re having status epilepticus they should be loaded with some sort of anti-epileptic agent. Even if they are chronically taking one, then either they are nonadherant or the drug isn’t working. A reasonable approach might be: – If nonadherance seems likely based on history, consider re-loading them on their chronic medication (this might make for an easier transition to chronic PO therapy). – If it seems like they were taking their medication correctly, consider loading them with a different anti-epileptic… Read more »
That makes a lot of sense to me, Thank you for you kind and quick answer
Assuming u r already in a controlled setting, do u actually wait for 5 minutes till u give benzos, what’s the harm in giving it immediately if the patient started to seize?
Can this algorithm be applied to pediatric patients in status epilepticus as well? If not, what other medications should i consider using?
Hello doctor,
What dose would you use for ketamine (1mg/Kg to 2mg/Kg iv push)? Any role for low-rate ketamine infusion?
How much time prior to the realiable EEG recording should sedatives be turned off? Do you have any articles on that theme, i cannot find any.