Non-convulsive status epilepticus

"Seizure disorders including status epilepticus" is a Level 1 topic from Section 2.1.8 in the first edition of the CICM Syllabus for the Second Part Examination, where the "including status epilepticus" presumably envelops all the categories of seizure disorder, including ones without classical tonic-clonic manifestations.  For whatever reason, the College loves non-convulsive status epilepticus. It has come up many times as one of the differentials. Notable occurrences where it was the chief topic of an SAQ include Question 20 from the second paper of 2012 and Question 4 from the first paper of 2024. 

• Nonconvilsuve status epilepticus  is a change in behaviour and or mental processes from baseline associated with continuous epileptiform EEG changes but without major motor signs.
• Epidemiology:  likely underdiagnosed, 8% of unexplained coma and ~16% of unexplained delirium in the elderly
• Aetiology: often associated with critical illness, especially neurological, e.g. hypoxic/ischaemic coma, TBI, stroke, etc.
• Pathophysiology:  Distingusihable from from convulsive SE by not involving motor cortex neurons, but also may be distinct in other ways, eg. potentially not a distinct disease process but rather a marker of severity of CNS dysfunction 
• Clinical course and outcome: prognosis is determined mainly by the underlying disorder (which is often severe, making it difficult to attribute morbidity and mortality to the NSCE itself)
•  Assessment: 
  • history of risk factors, mostly structural brain abnormalities or a history of seizures and an acute infection
  • characteristic examination findings of confusion, echolalia, automatism, nystagmus, mutism, facial twitching, etc 
  • EEG is gold standard, but one typically also asks for standard bloods and neuroimaging
• Management
  • Controversy exists as to whether one should offer any treatment at all
  • Treatment of the underlying precipitant is essential
  • Guidelines are mostly extrapolated from convulsive SE
  • Benzodiazepines are first line
  • Valproate, levitiracteam and lacosamide are second line
  • Seek to reduce threshold lowering stimuli (eg. hyponatremia)
  • Antiepilpetic therapy must be guided by serial or continous EEG, as clinical features are subtle and difficult to track

Oh's Manual explores this this issue in Chapter 49 (pp. 549, "Disorders  of  consciousness")  by Balasubramanian  Venkatesh, and Chapter 50 (pp. 560, " Status  epilepticus"  by Helen  I  Opdam.

Additional (free to read) resources could be scraped up from the internet:

• A good review from the Lancet (2007)
• An older BMJ observational study (1996)

Long & Koyfman (2023) are ostensibly writing their review "for emergency clinicians" but it is quick and to the point, which is a desirable characteristic for the CICM exam candidate who can no longer depend on remembering past exam questions.  The reader with severe OCD traits will revolt against (or insist vehemently on) the use of the hyphen for non-convulsoive status epilepticus, but the use of the unhyphenated version is ubiquitous and fails to confuse even the ever-literal DuckDuckGo, so it clearly doesn't matter, and the rest of this chapter will use both versions interchangeably and randomly. 

Definition of nonconvulsive status epilepticus

Borrowing from the fine tradition of the First Part exam, a definition is probably a good way to start any discussion, and for NCSE this might seem fairly straightforward (continued seizures without convulsion, innit). Unfortunately there is a whole range of different ways to say this exact same thing, which could mean that the exam candidate could mistakenly quote the wrong synonym lose marks. It seems that only one professional organisation has made an effort to create some standards here:

"Any seizure or status epilepticus without prominent motor activity" 

Or, in an already comatose patient:

"seizure activity for over 10 continuous min or over 20% of 1 h (12 min) on EEG monitoring"

- ACNS (2021)

... "Those forms with prominent motor symptoms and impairment of consciousness may be summarized as convulsive SE as opposed to the nonconvulsive forms" - ILAE (2015)

"Electroclinical" and "electrographic" are additional ACNS nomenclature items that define NCSE that is detected only by EEG, vs. NCSE which has some kind of detectable clinical signs. The International League Against Epilepsy have an even more granular definition (as they do for everything else to do with epilepsy) which is so far beyond the realm of pragmatic ICU training that it does not bear repeating even in this increasingly verbose and irrelevant resource. A CICM-exam-ready definition that is good enough for government work can be extracted from the colleges' own remarks on  Question 20 from the second paper of 2012, where the trainees were expected to define it as

"Change in behaviour and or mental processes from baseline associated with continuous epileptiform EEG changes but without major motor signs"

It is impossible to track this definition back to any specific paper or guideline, suggesting that it is a CICM original; but it is possible to track it forward to a paper by Ramesh Yasam et al (2017) who appear to have plagiarised it word for word without mentioning CICM. 

ontrol of seizures, prevention of secondary injury, finding and treating precipitants and involved a strategy for de-escalation

Assessment of suspected non-convulsive status epilepticus

Suspected, because that is how it usually arises- an entry into the list of differentials to explain why a patient may have an altered level of consciousness. Not necessarily depressed, mind you - "altered" is exactly the word one should use, because it is a "change in behaviour and or mental processes", which is nothing quite as crude as coma. But it can also be coma. 

History is a matter of building a case for proceedign the other (potentially expensive) NCSE investigations by collecting enough risk factors and predisposing features.  A study from Belfast reported on some risk factors identified among a series of patients with NCSE.  It should be added that even though the college lists these risk factors in Question 20, in fact this is from a well-established association. From over 1800 inpatient EEGs, there were 50 episodes of NCSE identified, and from this substrate the following risk factor profile was produced:

• Known epileptic with an acute infection
• Recent convulsive seizures
• Remote risk factors for seizures:
  • previous stroke
  • tumour
  • neurosurgery
  • dementia
  • multiple sclerosis
  • encephalitis or meningitis

To this we can add:

• Otherwise unexplained decreased level of consciousness
• New onset of psychosis
• History of substance abuse (especially a depressant, resulting in a depressed seizure threshold in withdrawal)

Though the defensive Americanised version of medicine presented by UpToDate suggests that "any fluctuating or unexplained alteration in behavior or mental status warrants consideration of NCSE and evaluation with EEG", practically speaking no sane person would subject every delirious patient to EEG. On the other hand, one should think carefully about non-convulsive status epilepticus in any patient in whom the decreased level of consciousness has not been explained by higher-yield investigations. Statistically speaking, one study has dicovered that among patients with undiagnosed coma, roughly 8% have the characteristic epileptiform discharges on their EEG.

Examination findings  are often nonspecific and subtle. In one study of clinical features of NCSE, "none of the clinical features evaluated had a high sensitivity and specificity individually. However, the combined sensitivity for “remote risk factors for seizures” and “eye movement abnormalities” was 100%." Remote risk factors in this situation included previous stroke, neurosurgical intervention, brain tumour, history of meningitis, and so on (as listed above).

Negative symptoms: anorexia, aphasia/ mutism amnesia catatonia confusion, coma lethargy fixed-gaze staring.

Positive symptoms: agitation/aggression automatisms uncontrollable blinking delirium, delusions, psychosis echolalia facial twitching (particularly, small periorbital muscles) nystagmus/eye deviation

Lattanzi et al (2021)

Investigations  for NCSE must include an EEG with epileptiform discharges, as this is necessary to confirm the diagnosis. Another helpful hint is the paradoxical restoration of normal alertness following the administration of an IV benzodiazepine. Other tests should exclude the usual suspects that cause delirium, which would include renal failure and liver disease. Thus:

• FBC to screen for macrocytsos as a clue to chronic alcoholism
• LFTs to exclude liver disease
• EUC to exclude renal failure
• CMP to exclude hypercalcemia
• CT brain to exclude a brain abscess or something equally obvious as the cause of the NCSE (though this does not so much diagnose NCSE, but rather explains it)
• Lumbar puncture
• MRI brain

Biomarkers are possibly disappointing, as one may find that the test is so exotic that it gets sent away to a distant private laboratory, from which result will be available perhaps weeks after the patient has already died.  Extrapolating from convulsive SE, Hanin et al (2020) recommended the following exotic tests:

• Neuron-specific enolase which leaks from injured neurons
• neurofilament light chain (NFL) which leaks from inured axons
• CSF acidosis
• increased CSF protein and CSF/serum albumin ratio
• S100 protein in CSF
• Glial fibrillary acidic protein (GFAP) in CSF

I want to do an EEG but the patient requires constant boluses of sedation for their agitation. Indeed, comrade, we have all been there. The medications used to control behaviour will also obscure the underlying seizures and produce an unhelpful report of "diffuse cerebral dysfunction" following EEG analysis by an expert. The solution is to use agents which have no effect on the seizure threshold and which have no intrinsic antiepileptic effects. Of these, the most attractive is dexmedetomidine, which generates an EEG pattern that resembles Stage II of sleep.

Difficulties in diagnosis of non-convulsive status

This became reworded as "challenges" instead of "difficulties" in Question 4 from the first paper of 2024, which implies that the examiners have come to view the diagnosis of NCSE as a fun opportunity for personal growth that tests your abilities and determination. Yes, but it wouldn't be a challenge if it wasn't difficult, for the following reasons:

• Non-specific clinical features.
• No agreement on the diagnostic criteria by which to identify this condition.
• The most frequently agreed upon criteria are non-clinical (i.e. they rely on EEG to identify epileptiform activity)
• Even when EEG is available, EEG findings are frequently non-specific which should also be interpreted as features of encephalopathy ( a 2012 guideline suggests that even vaguely rhythmic suspicious-looking EEG findings should be treated as NCSE).

Epidemiology of nonconvulsive status epilepticus

It would be wonderful to present something sensible here, but the only thing one can securely say is that NCSE is present in more patients than we believe. This can be said on the basis of the findings that:

• Among patients with undiagnosed coma, roughly 8% have the characteristic epileptiform discharges on their EEG.
• Among the elderly, Cheng (2014) claims it accounts for 16% of all causes of confusion without a specific aetiology, and this number seems to be repeated a lot by people who quote their paper, tthough from the reading of it one cannot be sure as to how this value was calculated.

In short, it is probably lurking in your nearby elderly ICU patient who is not completely normal and who has not yet dropped their level of consciousness to the degree that would typically arouse one's suspicion. 

Pathophysiology of nonconvuslive status epilepticus

Without overextending oneself into the lairs of venomous neurologists, one could safely remain in one's lane by referring the reader to experts like Rüegg (2017). It is also worth mentioning, from a superficial reading of the literature it appears that noncolvulsive status differs from the usual convulsive garden-variety status epilepticus mainly because it spares the primary motor neurons, i.e. it just so happens to be occurring away from the motor cortex.

Clinical course of non-convulsive status epilepticus

To be having continuous seizures is clearly unhelpful even when it does not cause the loss of airway reflexes. For example, Mesraoua et al (2017) followed 65 patients with NCSE and found that their mortality was a staggering 31%, putting it in the same category as septic shock and ARDS. On the other hand, this makes perfect sense, as NCSE does not typically happen for no reason in healthy people, and the cohort of patients sampled for this sort of study often have serious problems (in the case of Mesraoua et al, more than half had either a head injury, or a stroke, or were recovering from cardiac arrest). Smilarly, the 2012 Guidelines for Management etc etc contain within them a huge table (Table 4) on Page 7, which patiently lists the mortality statistics for all sorts of status epilepticus, and which all looks terrible:

Mortality: At hospital discharge: 18-52% At 30 days: 65% Time-critical diagnosis: Diagnosis within 30 min of seizure onset: mortality 36 % Diagnosed 24 hr after seizure onset: mortality 75 %

Features associated with poor outcome: Severe mental status impairment Longer seizure duration Less than 10hrs: 10% mortality More than 20hrs: 85% mortality Unknown cause

Given the difficulty in separating the effects of the disease from the effects of the status, we end up resorting to animal models, which demonstrate all kinds of neuronal damage folloiwbng prolonged seizure activity (Rajasekaran & Goodkin, 2018). Still, these authors were also ultimately forced to conclude that the prognosis of this condition mainly depends on the prognosis of the underlying disease process that caused it, and whether it has direct neurotoxic consequences in humans remains to be seen. Kaplan (2000) even went so far as to say that "le traitement peut se révéler pire que la maladie", implying that perhaps the toxicity of the high dose antiepileptic agents is doing some of the observed harm, and that perhaps the EEG findings are not the disease but rather an epiphenomenon which merely serves to call attention to CNS dysfunction of a certain severity.

Management of non-convulsive status epilepticus

A "discuss" question in the CICM exma vocabulary calls for an approach that describes the advantages, disadvantages and controversies in a subject, and this topic would be attractive for such an SAQ because it lends itself well to the discussions of uncertain benefits. It is actually not clear whether we should treat these people aggressively or not. One might summarise as follows:

Rationale and background

• Convulsive status epilepticus has nontrivial risks due to the acidosis, respiratory failure, aspiration risk, and other such phenomena
• Status epilepticus is also associated with some clearly defined neurotoxicity ("excitotoxicity") which can result in functional disability
• Nonconvusive status epilepticus does not clearly demonstrate a tendency towards either of these outcomes
• Ergo, one might make the argument that it should not be treated aggressively

Advantages of aggressive treatment:

• First line antiepileptic agents have low toxicity (so why not treat?)
• Poor outcomes are independently associated wth NCSE, and other features (eg. neuron-specific enolase leak) suggest that there is neurological damage occurring
• Seizures increase cerebral oxygen demand, which could contribute to the raised ICP of TBI patients

Disadvantages of aggressive treatment

• The association between neurological injury in NCSE is mostly based on animal models
• First line antiepileptics may not work, as the underlying condition is what drives the NCSE, and one may end up climbing a ladder of increasingly more and more toxic polypharmacy without much effect
• There is not much data to guide specific management for this condition, and we end up using potentially inappropriate methods, borrowing from the repertoire of therapies for the much more dangerous "normal" convulsive status epilepticus

Balance of pros and cons:

• Patients with life-threatening conditions would probably benefit most from the advantages, and would be the least affected by the disadvantages, where persistent EEG seizure activity is detected and the patient remains comatose
• Patients with mild absence type symptoms, or those who otherwise resemble a confused person with nothing else wrong with them, should probably be trialled on some small conservative dose of benzodiazepines, and if their response is unsatisfactory, one should have a low threshold for giving up and going after other low hanging fruit.

But say you decided to treat anyway. Which agents would be the best? We seem to extrapolate most of what we do here from normal SE.  The 2012 Guidelines for Management, referred to extensively in the summary of convulsive status epilepticus, do not offer anything specific for this condition. A different article from Clinical Neurophysiology (2007) also can't add anything specific. To form the current recommendations into a pointform list:

• A benzodiazepine  (does not seem to matter which) is first line
• Sodium valproate is next in line, it seems because it is effective for absence-type seizures, and some authors (eg. Lebedyn, 2020) suggest that it is a good choice because it is a "broad spectrum" antiepileptic
• Levitiracetam, lacosamide or phenytoin are often listed alongside valproate as first line, without any suggestion as to which is superior, which makes sense because nobody actually knows.
• Anaesthetic agents such as barbiturates are listed as the next in line, which resembles the management of refractory status epilepticus, in which case one must have convinced oneself that the NCSE is important enough to warrant a long and painful ICU stay.