A Case:
A 64 yo woman with history of COVID-19 and recent knee surgery presents to the ED obtunded. She is immediately intubated for inability to protect her airway. Non-con head CT is unremarkable. Given a concern for non-convulsive status a continuous EEG (cEEG) is ordered.
You receive the report:
This is an ABNORMAL continuous EEG:
- Continuous generalized and multifocal discharges and frequent polyspikes, occasionally with overriding fast activity (+F) and rarely with rhythmicity (+R) (within bursts).
- Burst suppression pattern at beginning of epoch with 2-4 seconds of burst intermixed with generalized epileptiform spike wave discharges up to 1-2 Hz and 2-8 seconds periods of suppression. 4 hours into the recording, interburst intervals shorten, and background becomes discontinuous theta/delta with continuous generalized epileptiform spike wave discharges, 1-3 Hz still with overriding fast and rhythmicity. These meet criteria for ictal-interictal continuum (IIC), and at times appear more on the ictal end of IIC. Reactivity was absent.
WHAT. IS. THAT.
EEG reports are filled with important information about the health of a patient’s brain, which can be used to direct treatment and prognosticate.
Many reads will come with a little “clinical significance” summary, but most epileptologists aren’t the ones caring for the ICU patient, you are. cEEG is one of the only continuous non-invasive modalities that can provide a lot of detail about the neurologic function in disease and health. And, interestingly, most seizing patients are taken care of in non-NeuroICU settings. In one study, for every one status epilepticus patient admitted to the neuro-ICU, three more were being cared for in the medical ICU (PMID: 23589183). Plus, as the diagnostic yield of EEG is highest when in close proximity to a witnessed event, EEG is a valuable tool in the ED (PMID: 21511976).
Thus, understanding the basics of these reports is important for all clinicians caring for acutely ill patients.
This post will cover:
👉 When to order continuous+/-video EEG (cEEG/cvEEG) monitoring
👉Some basic EEG terminology
👉 A review of the “Not Quite Seizures” Patterns & The “Ictal-Interictal Continuum”
Part I: When to order cEEG in critical care?
While the most obvious reason for cEEG is diagnosis and treatment of status epilepticus, there are many other reasons for ordering cEEG in critically ill patients:
Seizures, particularly, non-convulsive seizures are not uncommon among critically ill patients even without an underlying structural brain injury or without a primary neuro diagnosis:
Why not just use a routine EEG?
- Just like with a fib monitoring after a stroke… the longer you look, the more you will find.
- In this study, among comatose patients, 87% of seizures were captured by the first 48 hours, although another 9% seizures required 5 more days of monitoring.
But cEEG is labor intensive and not available everywhere…
- Correct
- The 2HELPS2B score is a validated score based mostly on EEG features in the first 1 hour of monitoring, and then stratifies patients into varying durations of monitoring to detect 95% of seizures within the next 72 hours (PMID: 31930362).
- Note that although most of the 2HELPS2B score is determined by how the EEG looks, you get 1 point for a history of epilepsy or recent clinical events concerning for seizures.
- Patients with a score of 0 can be disconnected from EEG after the 1 hour screen; patients with a score of 1 are recommended to have 12 hours of monitoring (so that's any patient with a recent clinical event that was concerning); and those with a high score (=> 2) should have 24 hours for seizure detection.
- You can find the 2HELPS2B calculator here
Part II: EEG Basics
EEG is most often obtained with scalp electrodes in a “10-20” arrangement with even electrodes on the right, odd electrodes on the left.
The dipole between two electrodes is displayed as a channel and the channels can be arranged into montages that allow comparison between different areas of the brain. How to read an EEG is beyond the scope here (obvi), but the terminology you get back is largely evaluating these 4 categories:
1️⃣ Background
2️⃣ Reactivity
3️⃣ Seizures
4️⃣ The panoply of everything else:
⚡️Sporadic discharges
⚡️Rhythmic and Periodic Patterns
⚡️Ictal-Interictal Continuum Patterns
💥Background
Brainwaves have frequencies and (unfortunately) those frequencies were not named in order of fastest to slowest (which would have made sense), but in the order in which they were discovered. The frequency is counted by how many oscillations there are in 1 sec (the vertical lines in the EEG clips below).
(above are frequencies for older children and adults; after age 8 the normal resting background reaches 8Hz – 8 for 8!)
Healthy, Normal EEG:
Remember Alpha = 8-13 hz, Beta > 13 hz
By comparison, the “typical” ICU EEG has a much slower background composed of theta and delta slowing and may have features of sharp or rhythmic patterns (more on those later):
A particularly ominous background is the “burst suppression” background, which (as the name suggests) describes an EEG that alternates between periods of complete suppression interspersed with “bursts” of EEG activity. This EEG pattern, which is deeper sedation than general anesthesia, may be induced with high levels of anesthetics OR may occur spontaneously after hypoxic ischemic injury or severe brain injury where it generally portends a poor prognosis.
💥Reactivity
Next an EEG is examined for reactivity. Reactivity describes a change in the EEG background resulting from an audial or tactile stimulus. Reactivity within the first 24 hours is a good prognostic sign after cardiac arrest. Those with reactivity have a 3-8x higher rate of obtaining a good outcome (PMID 32651293). Lack of reactivity has less prognostic significance.
💥 Seizure
Next the EEG is reviewed for any seizures.
Per ACNS terminology seizures have a strict electrographic definition:
- Electrographic Seizures are defined briefly as epileptiform discharges averaging >2.5 Hz for >= 10 seconds or any pattern with definite evolution lasting >= 10 seconds
- ⚡️Said simply, a seizure typically has a spike-wave patten that recurs faster than 2.5 repeats / seconds and continues for 10 seconds. Seizures have an onset and resolution; they evolve over the seconds they last.⚡️
- Electroclinical Seizures require a definite clinical event
Here is a nice example of an electroclinical seizure:
This is not subtle. Some seizures are much more subtle and take a lot more finesse to spot. In general, this is binary though: seizures, yes ✅ or no ❌ .
If seizures, treat seizures.
Status Epilepticus also has an “electrographic” and “electroclinical” definition, which requires 10 mins of continuous electrographic seizure or 20% of an hour long reading. Clinical, as you guessed, means there is clinical correlate.
If status, treat status.
Part III: Decoding Everything Else
The tricky part becomes that when someone has an injured brain or encephalopathy, you’ll find all sorts of interesting spike waves (epileptiform discharges) or rhythmic and periodic patterns that don’t meet the definition for seizure but aren’t normal either.
Table of “Not Quite Seizures” Terms Frequently Mentioned in EEG Reports
BIRDs (brief potentially ictal rhythmic discharges):
Think of these as “mini-seizures” – they have the features of a seizure, but they just don’t quite get to the 10 second threshold. Depending on the clinical situation, these often merit treatment.
Epileptiform discharges:
Waves that have a “seizure-like” morphology but they occur sporadically without evolution or rhythmicity. The “+ modifier” signals that there is something “angrier” (ie more epileptogenic) about the way these waves appear. If these are rare and sporadic, they signal mild epileptic potential. However, as these pick up in frequency and + modifiers or become more rhythmic, the more worrisome they become. Once they get to the 2.5hz range and happen continually for 10 seconds… that’s a seizure.
Triphasic Waves:
Yes, you've probably seen this pop up in lots of the general ICU population. This is a specific epileptiform discharge morphology that is usually seen in toxic-metabolic encephalopathies (classically hepatic encephalopathy, uremia) and multi-organ failure. More recently, it’s been described in cases of cefepime induced neurotoxicity. Generally, this is more a marker of underlying brain dysfunction and less a marker of impending seizures.
The Ictal-Interictal Continuum:
This category includes rhythmic patterns that disrupt the background. They may contribute to poor mental status and perhaps contribute to neuronal injury. The four most mentioned are Generalized Rhythmic Delta Activity (GRDA), Lateralized Rhythmic Delta Activity (LRDA), Generalized Periodic Discharges (GPDs), and Lateralized Periodic Discharges (LPDs). There are others.
Generalized Rhythmic Delta (GRDA):
A non-specific pattern that is slow (<5 hz) and may be seen in profound encephalopathy, post-ictally or with inflammatory, degenerative, traumatic, or toxic-metabolic disorders.
Lateralized Rhythmic Delta (LRDA):
Another lateralized slow (<5hz) pattern that usually reflects the presence of a focal lesion; associated with risk of acute seizures, especially non-convulsive status epilepticus.
Generalized Periodic Discharges (GPDs):
Sharp waves or spikes that occur rhythmically in the right and left brain. They may occur infrequently (0.5hz) or quite frequently (2hz). Tend to be seen in diffuse processes: toxic-metabolic encephalopathy, sepsis. But may also be seen in herpes simplex virus encephalitis and autoimmune encephalopathies. Can be associated with NCSE
Lateralized Periodic Discharges (LPDs):
Lateralized sharp waves or spikes which made have associated slow waves. Also can be seen infrequently or frequently (2hz). Commonly encountered in stroke, intracerebral hemorrhage, subarachnoid hemorrhage, tumors, abscesses, Creutzfeldt-Jakob disease, herpes simplex virus and other infectious/autoimmune pathology. LPDs are highly associated with seizures, especially in the setting of acute illness, metabolic disturbances or focal lesions.
This graphic is helpful in defining which patterns are likely to evolve into seizures:
From PMID: 27992625
As you can see here, generalized rhythmic delta has a very low risk of evolving to seizures despite the frequency, while LPDs have a significant risk despite the frequency. LRDA and GPDs have a significant risk only when they start to be more frequent than 1.5 hz.
Is the IIC pattern causing harm?
Still debated, but based on PET data, it does seem that there are patterns that result in hypermetabolism, thus increasing the cerebral metabolic rate, which may cause neuronal injury. This “heat map” of ictal-interictal continuum findings can help contextualize the findings reported in the EEG report:
SB = suppression burst (another way to say burst suppression). RDA = rhythmic delta activity (includes LDRA and GRDA). GPDs and LPDs as defined above. SW = sharp waves (another term for epileptiform discharges)
Then, on the far right, you have the clearly ictal patterns: generalized convulsive status epilepticus (GCSE), nonconvulsive status epilepticus (NCSE), and epilepsia partialis continua (a form of focal status epilepticus with ongoing focal motor seizures, typically felt to cause less brain injury given its focal nature).
The color coding above reflects cerebral metabolism on PET scan, with red signaling hypermetabolism on PET and blue representing hypometabolism.
From PMID: 27169855.
Finally, are these patterns what is causing my patient to not wake up?
Maybe. An approach to management of the interictal continuum really deserves its own post, for now here are a few papers that summarize the different approaches to management:
With this knowledge, back to the case:
This is an ABNORMAL continuous EEG:1) Continuous generalized and multifocal discharges and frequent polyspikes, occasionally with overriding fast activity (+F) and rarely with rhythmicity (+R) (within bursts). 2) Burst suppression pattern at beginning of epoch with 2-4 seconds of burst intermixed with generalized epileptiform spike wave discharges up to 1-2 Hz and 2-8 seconds periods of suppression. 4 hours into the recording, interburst intervals shorten, and background becomes discontinuous theta/delta with continuous generalized epileptiform spike wave discharges, 1-3 Hz still with overriding fast and rhythmicity.
These meet criteria for ictal-interictal continuum (IIC), and at times appear more on the ictal end of IIC. Reactivity was absent.
Breaking this down:
- Background = Burst suppression pattern = Pattern caused by anesthetics or severe brain injury. Considered a malignant pattern if seen spontaneously (eg in hypoxic-ischemic encephalopathy). A very abnormal background.
- Reactivity = Absent.
- Epileptiform Patterns = Bursts have generalized epileptiform discharges with some “+” modifiers and are at times 2 hz.
- This is telling you these angry bursts look like seizures, but don’t quiet meet criteria for seizures
- Remember – 2.5 hz for 10 seconds = seizures. What is described is 2hz for 2-4 seconds; still angry but not quite the definition of seizures.
Taken together: this is a highly abnormal background which likely is indicative of a drug effect or severe brain dysfunction. The bursts look very spike-y, like seizures, as they are rhythmic and fast but only last 2-4 seconds. This is a very concerning pattern and is alerting you to a malignant, “nearly-seizure” EEG = “[the bursts are on the] ictal-interictal continuum (IIC), and at times appear more on the ictal end of IIC”
This looks like:
This is absolutely a pattern that is worth a treatment trial.
Ultimately, this case was of a patient who presented with an (unknown) overdose of baclofen. She came in in burst suppression because of the anesthetic effect of baclofen toxicity, and then, as she went through withdrawal her EEG became less and less suppressed and more and more on the ictal end of the interictal-ictal continuum, ultimately culminating in NCSE that was aggressively managed. She recovered. In this case, it was the way the EEG changed overtime, coupled with a negative imaging workup for hypoxic ischemic damage, that clued us in to a potential overdose and withdrawal situation, which was eventually confirmed with family.
Summary:
Some take-away pearls for EEG interpretation:
Photo by israel palacio on Unsplash
- EMCrit 336 – Team NeuroEMCrit's Critical Neuro Cases – Part 2 - November 3, 2022
- NeuroEMCrit – Team NeuroEMCrit's H&R Conference Talk, Part 1 - October 3, 2022
- NeuroEMCrit – The Perils of the ICH Score - August 1, 2022
Simply A – M – A – Z – I – N – G ! Thank
Thank you!
I am very sorry to admit this but I got my MD in ’82 (1982)… and I never understood EEG’s. as a matter of fact, not sure it was ever taught. they were always this mysterious cloud, hard to order, harder to get, and even harder to interpret. thanks to Casey and Neha, a few rays of sunshine are peeking through the beta and theta clouds of my mind regarding this important tool. over the years, I’ve always wondered which of my post-ROSC comatose patients were in NCSE, especially in the non-tertiary ER’s/ICU’s. I wonder if the Cerebell device and… Read more »
Wow, I’m so happy this was that helpful! thanks for the feedback 🙂
Truly awesome as per usual. I would add that in the era of ubiquitous CT perfusion you can use the opportunity of having your acutely obtunded pt in the scanner anyway for rule out of badness to address status as well as LVO (takes less contrast medium; in our institution just 40 ccs compared with 60 ccs for CTA, still the corcle is visualized fine). Although not prospectively studied, CTP seems to be more sensitive than eeg (can see temporomesial hyperperfusion where eeg can only reach with sphenoidal needle electrodes) and way easier and faster. Of course this only substitutes… Read more »
Yes absolutely! CTP has a lot of potential here to assess for status, but of course as you point out is only one time point — but certainly has potentially to be part of the initial workup. Brain PET/SPECT scanning really shines here, but SO hard to get, maybe one day it will be easier!
Thank you so much for this explainer! (Anecdotally) Baclofen overdoses seem to be increasing and as a result cEEG reads increasingly appear in our charts. This provided a lot of clarity to an area pharmacists don’t typically explore in didactic/experiential education.
Simplified & demystified, outstanding
Its clearly study material for medical students, it’s good that you had shared your experience.
by Raghotham from 4 Pole MCB