Introduction

Continuous electroencephalogram (cEEG) is a noninvasive tool that provides a unique opportunity for monitoring brain activity in real-time manner. The primary indication of cEEG is the detection of nonconvulsive seizures and the evaluation of paroxysmal events; its use has now expanded to include ischemia detection in selected neurocritically ill patients and in neuroprognostication, as recognized by the American Clinical Neurophysiology Society (ACNS) [1]. The widespread use of cEEG has allowed for the recognition of the frequent occurrence of electrographic seizures and of rhythmic and periodic patterns in critically ill patients. Determining the ictal versus interictal nature of these patterns and developing the next steps in the plan of care of such patients can be quite challenging in practice.

The ACNS has proposed a unified nomenclature for describing rhythmic and periodic patterns, aiming at creating a universal language for defining those patterns and streamline research [2]. A summary of main terms and modifiers included in the most recent version of this nomenclature is displayed in Table 1. Patterns are described based on their location (main term 1) and type of recurring discharges (main term 2). The three main pattern types include recurring discharges of relatively uniform morphology and are divided into periodic discharges (PD), rhythmic delta activity (RDA), and spike-and-wave or sharp-and-wave discharge (SW). In PDs, the discharges occur at nearly regular intervals with a clearly defined inter-discharge interval, whereas in RDA and SW, there is no intervening inter-discharge interval. The SW discharges are formed of spike, polyspikes, or sharp waves, immediately followed by a slow wave in a regularly repeating and alternating manner (e.g., spike-wave, spike-wave, spike-wave). Multiple modifiers can be used to characterize other features of the pattern of interest. The interrater agreement (IRA) for describing these patterns was nearly perfect for main terms 1 and 2, moderate or substantial for most modifiers, and fair for evolution according to one recent study [3]. The fair IRA for evolution can be partially explained by the provision of short epochs in which evolution can be difficult to appreciate in this study.

Table 1 Standardized critical care EEG terminology for rhythmic and periodic patterns
Full size table

There have been multiple proposed sets of criteria for defining electrographic seizures and nonconvulsive status epilepticus (NCSE). The current and most widely used criteria were initially proposed by Young et al. in 1996 and have gone through multiple revisions since the original publication [4]. The most recent revision was proposed at the 4th London-Innsbruck Colloquium on Status Epilepticus in Salzburg, 2013, and is commonly referred to as the Salzburg criteria (Table 2) [5]. Based on these criteria, epileptiform discharges occurring at a rate of more than 2.5 Hz and epileptiform or rhythmic discharges associated with typical spatiotemporal evolution or subtle clinical phenomena are considered ictal patterns (i.e., NCSE). However, there are a variety of rhythmic and periodic patterns with a high degree of uncertainty regarding their ictal nature. These present a diagnostic and therapeutic dilemma to electroencephalographers, intensivists, and general neurologists taking care of these patients. The term “ictal interictal continuum” (IIC) was first introduced by Pohlmann-Eden et al. in 1996, who described periodic lateralized epileptiform discharges (formerly called PLEDs, now referred to LPDs according to the new ACNS nomenclature) as “an electrographic signature of a dynamic pathophysiological state in which unstable neurobiological processes create an ictal interictal continuum, with the nature of the underlying neuronal injury, the patient’s preexisting propensity to have seizures, and the coexistence of any acute metabolic derangements all contributing to whether seizures occur or not” [6]. The use of the term has now expanded to include other rhythmic and periodic patterns that are not definitely ictal, but could still be, and that may contribute to neuronal injury in certain clinical settings. There is no consensus agreement on IIC patterns definition, but these generally include rhythmic and periodic patterns occurring at a rate of 1–2.25 Hz―particularly when fluctuating and including “plus modifiers” (Table 1). In this review, we shed some light on the data available on the significance of those patterns and the approach to managing patients harboring them.

Table 2 Salzburg criteria for electrographic seizures and nonconvulsive status epilepticus
Full size table

Case report

A 24-year-old man was admitted to the hospital following a witnessed first lifetime generalized tonic-clonic seizure while standing in line at a restaurant. According to the family, he was experiencing forgetfulness and difficulty maintaining concentration during the preceding weeks. They denied any mouth or hand automatisms, behavior or speech arrest, or limb shaking. On admission, he received 2000 mg of levetiracetam and was started on 750 mg every 12 h as maintenance therapy. A brain magnetic resonance imaging (MRI) demonstrated a cortically based right frontal 3.1 × 2.3 × 3.0 cm mass with peripheral and internal nodular enhancing foci for which he underwent an uneventful gross total resection (pathology consistent with World Health Organization grade III anaplastic oligoastrocytoma) 2 days later. Immediately following surgery, he was noted to have mild left hemiparesis, which gradually improved. On post-operative day 1, he experienced severe nausea and was noted to be more somnolent. A repeat noncontrast head CT demonstrated expected post-surgical changes with some blood in the surgical cavity. Given the concern for subclinical seizures possibly related to post-operative irritation and edema, levetiracetam dosing was increased (received a total of 2500 mg in 12 h) and 1 g/Kg of mannitol was administered; both interventions led to no change in the exam. He subsequently developed aphasia and on the morning of post-operative day 2, he was connected to continuous EEG, which demonstrated left temporal fluctuating lateralized rhythmic delta activity, at times spreading broadly throughout the left hemisphere with embedded spikes (LRDA+S) reaching 2 Hz. Since there was no clear onset/offset to suggest unequivocal seizures, nor was the pattern sustained at > 2.5 Hz to qualify as NCSE, this pattern was on the ictal end of the IIC (Fig. 1a–c). Given the presence of this potential ictal pattern despite a therapeutic dose of levetiracetam (3000 mg/24 h), a second agent was added. Approximately 45 min following a 15 mg/kg load of phenytoin, there was a significant electrographic response to treatment with a slow and gradual clinical improvement over the course of hours (Fig. 1d, e).

figure 1figure 1

This case has many notable points: (a) a clinical response to anti-seizure therapeutic trial seals a diagnosis of NCSE despite the fact that the electroencephalographic pattern did not meet criteria for unequivocal NCSE (sustained discharges > 2.5 Hz); (b) the clinical, and electroencephalographic, response to a therapeutic trial may be gradual over hours; (c) the fact that the patient did not respond to adequate doses of levetiracetam (although the patient was not monitored on EEG when this drug was given) does not rule out the potential ictal nature of his symptoms; (d) the epileptogenic focus may be remote from any obvious structural abnormality: in this case, the patient had a right frontal tumor resected (which is an obvious potential epileptogenic source); however, the epileptogenic focus was contralateral to his lesion.

Epidemiology

Rhythmic and periodic patterns have been reported to occur in 30–37% of patients undergoing cEEG monitoring [7, 8]. In these studies, patterns included lateralized periodic discharges (LPDs; previously called PLEDs), bilateral independent periodic discharges (BIPDs; previously called BIPLEDs), generalized periodic discharges (GPDs; previously GPEDs), lateralized rhythmic delta activity (LRDA), and generalized rhythmic delta activity (GRDA). Although less commonly seen, multifocal (Mf-) periodic discharges and rhythmic delta activity and any SW are also included in rhythmic and periodic patterns. LPDs are primarily seen in the setting of acute structural brain lesions, most commonly ischemic stroke [9, 10, 11], and tend to resolve over weeks following acute injury or seizure [12]. BIPDs most commonly occur in patients with structural brain injury as well, though they are more frequently seen in the setting of bilateral brain lesions [11]. On the other hand, GPDs are most frequently seen in the setting of metabolic derangements [13, 14, 15]. Other causes for GPDs include acute brain injuries [13, 14, 15], post-anoxic encephalopathy [13, 16], Creutzfeld Jacob disease (CJD) [15], subacute sclerosing panencephalitis (SSPE) [15], and toxic exposure such as cefepime [17, 18] and lithium [19]. LRDA, a recently described pattern, is seen most frequently in the setting of acute or remote brain injury [20]. Frontally predominant GRDA (GRDA; previously termed frontal intermittent rhythmic delta activity; FIRDA)―the most common form of GRDA―is considered a nonspecific pattern reflective of various degrees of encephalopathy [21]. Notably, most studies investigating this pattern predate the current ACNS terminology and the term “FIRDA” was more loosely applied to any rhythmic appearing frontally predominant delta frequency discharges; many of which would not fulfill the current ACNS criteria for GRDA. Stimulus-induced rhythmic, periodic, or ictal discharges (SIRPIDs)―initially described by Hirsch et al. in 2004―encompass rhythmic, periodic, or ictal-appearing discharges induced by alerting stimuli [22]. This term was replaced in the ACNS terminology by a modifier denoting whether the pattern was spontaneous or stimulus-induced (e.g., SI-LPDs, SI-GPDs, and so forth) [2]. Often, SIRPIDs are not associated with any clinical correlate, but there have been few reports of focal motor seizures induced by alerting stimuli [23] contradicting the common notion of stimulus-induced patterns being exclusively interictal in nature.

Significance

Various rhythmic and periodic patterns are associated with increased seizure risk. The incidence of seizures during the course of acute illness ranges from 45 to 85% among patients with LPDs [7,8,9,10,11, 20], 43–78% among patients with BIPDs [11, 24], 11–89% among patients with GPDs [7, 8, 13–15], and 35–63% among patients with LRDA on EEG [8, 20]. Meanwhile, GRDA―even when associated with a plus modifier―was not associated with increased seizure risk, based on one multi-center retrospective study including 927 patients [25••]. Thus, GRDA is usually not considered a pattern that lies on the ictal interictal continuum. In addition to their long recognized association with seizures, periodic discharges have been associated with worse short-term outcome in patients with poor-grade subarachnoid hemorrhage (SAH) [26], intracerebral hemorrhage (ICH) [27], and even in patients with no evidence of acute brain injury [28]. The significance of SIRPIDs remains controversial. SIRPIDs were not associated with in-hospital mortality after controlling for confounders in a recent multicenter study including 416 critically ill patients undergoing cEEG [29••], while another recent study reported SIRPIDs as a poor prognostic determinant in patients with post-anoxic encephalopathy, particularly when recorded during therapeutic hypothermia [30]. Of importance, one recent study demonstrated that IIC patterns were strongly associated with the development of delayed cerebral ischemia in patients with SAH [31]. Furthermore, a recent prospective study demonstrated an association of LPDs on EEG with an increased risk for long-term development of epilepsy [32]. It is important to note that a causal relationship cannot be drawn from these data, and it remains debatable whether IIC patterns represent an ictal or interictal phenomena, or merely an epiphenomena of underlying structural or functional neuronal injury. Consequently, there is no consensus on whether these patterns need to be treated as seizures or not, and how aggressive a clinician should be in treating those patterns if the decision is made to treat them at all. Further, to this date, data guiding the pharmacologic approach to patients with IIC patterns, comparing watchful monitoring with therapeutic trials, and exploring how aggressively therapeutic trials should be implemented are lacking. Treatment algorithms are largely based on experts’ opinion and extrapolated from studies with NCSE.

Arguments for the ictal nature of certain IIC patterns

  1. 1-

    Focal clonic movements―including epilepsia partialis continua―can be seen time-locked to LPDs (ictal LPDs) and represent brief seizures, particularly in those with fronto-central fields [33]. The lack of clinical correlate in patients with LPDs localized to other regions could therefore be due to involvement of “silent brain regions.” Negative motor and sensory phenomena including hemiparesis, aphasia, amnesia, apraxia, and cortical blindness have been reported in association with LPDs [34]; this further substantiates the evidence supporting this hypothesis. Clinical correlates in ictal LPDs are seizures and may resolve completely upon administration of anti-seizure medications [35] (although some cases may be refractory and hard to treat), denoting the ictal nature of LPDs in this setting.

  2. 2-

    There are multiple reports of clinical improvement coupled with EEG pattern resolution following intravenous (IV) benzodiazepine [36,37,38,39] or non-sedating anti-seizure drug [38] administration in patients with IIC patterns associated with disturbed level of consciousness. We will further expand on medication trials later in this article.

  3. 3-

    IIC patterns have been reported in association with known imaging markers of status epilepticus. These include restricted diffusion on diffusion-weighted imaging (DWI) on MRI [40], increased regional cerebral flow on computed tomography (CT) [41], or MRI-perfusion studies [40, 41] as well as single-photon emission computed tomography (SPECT) imaging [40, 43], and regional hypermetabolism on fluorodeoxyglucose (FDG)-positron emission tomography (PET) scans [44, 45]. FDG-PET hypermetabolism predicted electroclinical status epilepticus with 79% sensitivity and 100% specificity in IIC patients undergoing cEEG according to one recent study including 18 patients [45]. These changes often reversed with treatment in a similar fashion to what would be expected in the setting of unequivocal status epilepticus [40].

  4. 4-

    IIC patterns are frequently associated with markers of neuronal injury such as increased lactate/pyruvate ratio (LPR) and decreased glucose levels in cerebral microdialysate (a modality of invasive multimodal monitoring) [46]. Periodic discharges as well as nonconvulsive seizures were temporally associated with metabolic crisis—defined as increased LPR and concurrent focal neuroglycopenia—in patients with traumatic brain injury according to one recent study [46]. These findings raise the question whether these potentially avoidable (or treatable) findings could be targeted to mitigate the development of secondary neuronal injury in traumatic brain injury.

  5. 5-

    Periodic discharges have been reported as an intervening pattern in patients with unequivocal status epilepticus [47].

  6. 6-

    IIC patterns on scalp EEG can be associated with unequivocal seizures on simultaneous intracranial recordings using depth EEG. Claassen et al. analyzed intracortical EEG with simultaneous scalp EEG and multimodality physiological monitoring recordings, and demonstrated that up to 19% of seizures detected on depth EEG were associated with IIC patterns on scalp recordings [48].

Arguments for the interictal nature of certain IIC patterns

  1. 1-

    Rarely, chronic LPDs may be seen in patients with longstanding epilepsy in a similar fashion to other interictal patterns [49]. However, LPDs in this setting generally occur at a frequency of < 1 Hz, which would not fulfill our definition for IIC patterns. Nonetheless, one form of LRDA (temporal intermittent rhythmic delta activity—TIRDA), is a long-recognized interictal pattern seen in patients with temporal lobe epilepsy, particularly those with mesial temporal sclerosis[50]; this raises the possibility that their LRDA counterparts seen in ICU setting may represent similar interictal phenomena.

  2. 2-

    Electrographic seizures often emerge from periodic discharges similar to their emergence from other interictal patterns coupled with disappearance of periodic discharges, arguing for PDs being an interictal rather than an ictal pattern. Similarly, one study using electrocorticography concurrently with scalp cEEG demonstrated disappearance of LRDA on scalp (pseudonormalization) correlating with intracranial seizure recording [51].

  3. 3-

    Periodic discharges are not consistently associated with the restricted diffusion on DWI imaging that is typically seen in patients with status epilepticus. In one recent study including 10 patients with LPDs, all 5 patients with electrographic seizures in addition to LPDs had restricted diffusion in the region of LPDs, whereas all other 5 patients with isolated LPDs had no areas of restricted diffusion [52]. Apparent diffusion coefficient (ADC) changes on MRI have been strongly correlated with the degree of seizure-induced neuronal injury in animal studies [53], and therefore, the lack of diffusion restriction associated with LPDs may suggest that these are interictal findings rather than unequivocally ictal, or at least “less malignant” ictal patterns potentially warranting less aggressive management. Further, there have been few reports of patients with SIRPIDs lacking the regional cerebral hyperperfusion that is typically seen in patients with unequivocal seizure activity [54, 55].

Taken altogether, these arguments on both sides suggest that not all rhythmic and periodic patterns should be treated in the same way and that certain patterns are more likely to be ictal (or more harmful) and may warrant more aggressive management with pharmacologic therapy.

Suggested approach to patients with rhythmic and periodic patterns on cEEG

The following is our suggested approach to these patterns (summarized in Figure 2):

  1. 1-

    Exclude clear ictal or interictal patterns:

    Epileptiform discharges occurring at a frequency of ≥ 2.5 Hz and periodic discharges or rhythmic activity associated with unequivocal spatiotemporal evolution or subtle ictal clinical correlate are classified as ictal patterns [56, 57], and should be treated accordingly. On the other hand, most electroencephalographers agree that PDs or RDA occurring at a frequency of less than 1 Hz as well as static RDA without plus features are interictal. Patterns occurring at a frequency of 1–2.5 Hz, those with plus features, and/or displaying fluctuation are considered to lie on the IIC.

  2. 2-

    Identify EEG characteristics highly associated with seizures:

    Certain EEG features are strongly associated with seizures and, when present, represent patterns that lie on the IIC. These include distinctive morphological features of the pattern as well as higher pattern frequency. Reiher et al. in 1991 classified PLEDs (now called LPDs) into two main subtypes: PLEDs proper and PLEDs plus [58]. PLEDs proper included periodic patterns not associated with rhythmic discharges, whereas PLEDs plus included those with superimposed rhythmic discharges, most commonly in the form of low amplitude fast frequency discharges. Electrographic seizures were seen in 74% of PLEDs plus patients, compared to only 6% of patients with PLEDs proper [58]. In a more recent retrospective study investigating EEG characteristics associated with seizures in 100 patients with LPDs on cEEG, patterns with superimposed rhythmic activity “LPD + R” were associated with the highest odds for developing status epilepticus and seizures (OR 13.91; CI 5.3–36.52; P < 0.0001). In the same study, patterns with associated fast activity “LPD + F” were also highly associated with unequivocal ictal activity (OR 5.16; 2.08–13.07; P = 0.0004), whereas LPDs bearing blunt morphology were not (OR 0.24 (CI 0.097–0.62; P = 0.0036) [59••]. Generalized periodic discharges (GPDs) and LRDA were associated with seizures in a frequency-dependent manner, where only frequencies > 1.5 Hz were associated with increased seizure risk in one in one multicenter retrospective study including 4772 patients undergoing cEEG monitoring [25••]. Further, LPDs, GPDs, and LRDA with plus features were more likely to be associated with seizures than those without [25••]. Another study demonstrated that PDs > 2 Hz in frequency were associated a decrease in the partial pressure of oxygen in interstitial brain tissue (PbtO2); this relationship was dependent on the frequency of discharges: PDs > 2 Hz were more likely to be associated with secondary neuronal injury from lower regional oxygen saturation [60]. In summary, when periodic and/or rhythmic patterns that lie on the IIC are identified on cEEG, a careful clinical risk stratification is in order: patients with high risk for secondary brain injury or those who have evidence of ongoing neuronal injury may be the ideal candidates for a therapeutic challenge.

  3. 3-

    Correlate pattern with ancillary studies that reflect ongoing neuronal injury and/or may represent status epilepticus:

Fig. 2
figure 2

Clinical approach to patients with rhythmic and periodic patterns. Single asterisk indicates the evaluation for markers of neuronal injury may include serum biochemical markers (e.g., neuron-specific enolase), neuroimaging, and/or multimodal invasive monitoring. The access to these resources varies across different institutions. Double asterisks indicate anesthetic infusions have been associated with an increased morbidity; thus, these treatments are often reserved for unequivocally ictal patterns (nonconvulsive status epilepticus). In selected cases, these may be considered if further aggressive treatment is warranted. EEG electroencephalogram (cEEG continuous electroencephalogram), IIC ictal interictal continuum, PD periodic discharges, RDA rhythmic delta activity (LRDA lateralized rhythmic delta activity, MfRDA multifocal rhythmic delta activity, BIRDA bilateral independent rhythmic delta activity), SW spike-wave.

Full size image

Neuroimaging

One of the most useful approaches in investigating IIC patterns is to correlate the EEG with neuroimaging findings. Findings commonly seen in patients with status epilepticus include restricted diffusion on DWI imaging [52], increased regional perfusion on CT [41], MRI perfusion [40, 42] or SPECT imaging [40, 43], and hypermetabolism on PET scans [44, 45]. It is important, and particularly challenging, to distinguish these findings from those related to underlying illness; reversal of these changes following treatment further corroborates the hypothesis of them being a result of an ictal process.

Serum biomarkers

Elevation of serum markers of neuronal injury such as neuron-specific enolase (NSE) can often be seen in patients with SE, potentially serving as a marker of seizure-induced neuronal damage. Serum NSE levels correlated with severity of histological brain injury in a study using a rat model of status epilepticus [61]. Serum NSE levels generally peak 24–48 h after status epilepticus onset [62], regardless if nonconvulsive or convulsive, though focal status epilepticus with impaired consciousness (previously called CPSE) and subclinical status epilepticus are associated with the most significant NSE elevation [63]. Given the close association of NSE elevation with unequivocally ictal patterns, high titers in association with IIC patterns may provide evidence to justify their management as an ictal pattern given the potential for ongoing neuronal injury.

Multimodality monitoring

Recent data have highlighted the role of invasive multimodality monitoring as an emerging tool for continuous monitoring of physiological parameters that act as markers of ongoing neuronal injury, possibly providing a window of opportunity for early intervention before irreversible neuronal injury ensues. Commonly employed modalities include intracranial EEG monitoring via intracortical mini-depth or subdural strip electrodes, intracranial pressure monitoring via intraventricular catheters or intraparenchymal probes, monitoring cerebral tissue oxygenation (PbtO2) via specialized probes, indirect cerebral oxygenation via jugular bulb oximetry, continuous cerebral blood flow monitoring via thermal diffusion probes, and monitoring regional metabolism via cerebral microdialysis. Claassen et al. demonstrated that only 19% of intracranial seizures were associated with IIC patterns on scalp EEG [48]; these were also associated with increases in heart rate, respiratory rate, and mean arterial pressure (MAP), as well as trends towards higher intracranial pressure and cerebral perfusion pressure [48]. In addition, PDs can be associated with metabolic crisis, and PDs > 2 Hz can be associated with regional hypoxia (see above) [60]. This suggests that high-frequency (> 2 Hz) periodic discharges place an increased metabolic demand, which in turn may lead to bioenergetic mismatch in the setting of acute brain injuries and may lead to secondary neuronal injury. These markers when detected in association with IIC patterns may provide evidence for these patterns being on the ictal end of the IIC, thus warranting aggressive management in selected cases. It is important to remark that the use of these modalities remains largely investigational, and a causal relationship between IIC patterns and neuronal injury cannot be drawn from available evidence at this point.

Treatment approach

There is no consensus agreement on how to treat IIC patterns given the paucity of available evidence comparing the performance of different therapeutic modalities. Our suggested approach includes a careful consideration of the potential risk of secondary neuronal injury from increased metabolic demand based on the primary acute brain injury; thus, this is a case-by-case risk stratification. Often, we recommend primary seizure prophylaxis with a non-sedating anti-seizure medication using a standard maintenance regimen without a loading dose (e.g., levetiracetam 750 mg every 12 h). The escalation of pharmacologic therapy, including loading doses of the initial agent and the addition of a benzodiazepine trial, largely depends on the perceived ictal potential of the pattern characterized on EEG and its evolution over time, guided by the previously discussed risk factors and ictal correlates. Given the paucity of available data guiding management of those patterns, patterns thought to be on the ictal end of the IIC (see above) are usually managed in the same manner as NCSE. Table 3 provides a list of non-sedating and sedating anti-seizure medications used in NCSE, commonly used dosage ranges, and commonly encountered adverse events.

Table 3 Medications used in nonconvulsive status epilepticus
Full size table

A therapeutic trial is a simple, yet underutilized, diagnostic and therapeutic test that involves administering sequential small doses of short-acting benzodiazepines (e.g., 1 mg of midazolam), or alternatively, non-sedating anti-seizure medication (e.g., loading with fosphenytoin, levetiracetam, or valproic acid) to patients harboring IIC patterns, and monitoring for an electroclinical response. Resolution of the potentially ictal (IIC) EEG pattern coupled with clinical improvement or recovery of EEG background rhythms defines a positive response [64]. It is important to note that, at times, improvement may be delayed, which adds a level of complexity to these trials. Improvement of EEG alone is often termed “possible NCSE” [56]. Hopp et al. retrospectively analyzed EEG and clinical response to IV benzodiazepine trials in 62 patients with suspected NCSE and demonstrated positive clinical response in 35% of patients [36]. Positive clinical response correlated strongly with recovery of consciousness, survival, and with achieving a favorable functional outcome [36]. We often use non-sedating medications as IV phenytoin or fosphenytoin, levetiracetam, valproate, or lacosamide to avoid the confounding sedating effect of benzodiazepines, and appreciate a subtle clinical improvement. Recent data support the role of anti-seizure medication trial in diagnosing NCSE in patients with patterns on EEG that have triphasic morphology [38]. This is important, as these patterns have been historically considered as interictal in nature and reflective of metabolic encephalopathy, and their ictal potential is increasingly being recognized [63]. O’Rouke et al. investigated the clinical response to a therapeutic trial among 64 patients with TW pattern and demonstrated a positive clinical response in 10/53 (18.9%) patients treated with benzodiazepines and 19/45 (42.2%) patients treated with non-sedating drugs [38]. These data underscore the utility of IV therapeutic trial as a simple and inexpensive tool for diagnosing NCSE (an IIC pattern with a clinical response to medication trial). However, it is worth mentioning that there is no negative result to this test, as lack of EEG and clinical response does not completely exclude an ictal connotation (NCSE). The refractoriness of a potentially ictal pattern remains a possibility; thus, we recommend close monitoring and reassessments with either further diagnostic testing supporting neuronal injury or considering a trial with a different agent in these cases.

Non-sedating agents

Phenytoin is probably one of the most studied and frequently used medication for seizure prophylaxis in patients with acute brain injury, and in convulsive and nonconvulsive status epilepticus. In one of the largest prospective randomized placebo-controlled trials of seizure prophylaxis in severe traumatic brain injury, phenytoin prophylaxis was associated with a significant decrease in early seizures (within 7 days) but failed to demonstrate a decreased risk of late seizures (within 2 years) in this population [65]. On the other hand, data from animal [66] and human studies [67] linking phenytoin exposure to decreased functional and cognitive outcomes in stroke and subarachnoid hemorrhage have led to a decrease in its use. Furthermore, phenytoin’s unpredictable zero-order kinetics and unfavorable drug interaction profile hampering the efficacy of oral anticoagulants, antibiotics, and chemotherapeutic agents in addition to the associated risk of serious cardiovascular adverse events have limited its use in critically ill patients [68,69,70]. Nowadays, medications with more favorable pharmacokinetic profile (e.g., levetiracetam, lacosamide, brivaracetam) are being used more frequently. Valproic acid is another commonly used drug, but its associated risk of coagulopathy, hepatic dysfunction, and hyperammonemia [71] limits its use in certain circumstances. Other alternative therapies include carbamazepine, oxcarbazepine, topiramate, gabapentin, pregabalin, phenobarbital, brivaracetam, perampanel, clobazam, and vigabatrin.

Benzodiazepine trial

While sequential low doses of IV benzodiazepines (e.g., midazolam 1 mg in sequential doses up to a maximum dose of 0.2 mg/kg) have been utilized as a therapeutic bedside test to diagnose and potentially treat NCSE (see above), the role of continuous IV infusions in the management of IIC patterns resistant to first and seconds lines of therapy is less clear. The benefit of pattern suppression must be carefully balanced against the potential risk of iatrogenesis, and in the light of scarce evidence guiding therapeutic decisions in patients with IIC patterns, the risk/benefit ratio is largely extrapolated from status epilepticus studies. One study from 20 years ago demonstrated that IV benzodiazepine administration was associated with increased short-term mortality in critically ill elderly population with NCSE [72]. More recent studies have demonstrated that anesthetic use was associated with a poor outcome in patients with status epilepticus [73,74,75]. This association seemed to be stronger in patients with focal status epilepticus with impaired awareness [73]. However, none of these studies fully accounted for the severity and refractoriness of NCSE, nor did they assess long-term functional and cognitive outcomes in survivors, limiting the generalizability of these data. On the other hand, one retrospective study compared two cohorts of refractory status epilepticus treated with IV midazolam at different maximum doses. The first group (historical control) received up to a maximum dose of 0.4 mg/kg/h while the second group reached up to 2.9 mg/kg/h. Lower seizure risk and short-term mortality were seen in the second group, questioning the association of aggressive seizure management with poor outcome [76]. Randomized prospective studies are needed to solve this dilemma arising from conflicting evidence, but these studies are expensive and very difficult to conduct. Therefore, we favor the use of non-sedating anti-seizure medications as trials before resorting to anesthetic infusions, particularly in patients without a secured airway and with preserved consciousness, and in those with brief and intermittent focal seizures.

Conclusion

Ictal interictal continuum patterns present a diagnostic and therapeutic challenge to practitioners dealing with patients with acute brain injuries and critically ill patients. The presence of electrographic features suggesting a more likely ictal pattern or a positive response to a therapeutic trial can assist in the identification of those that lie on the ictal end of the continuum, and potentially carry a higher propensity for association with neuronal injury. Serum biomarkers, ancillary neuroimaging, and data from multimodality monitoring can provide further evidence for ongoing neuronal injury associated with IIC patterns, and may guide an individualized approach to patients harboring such patterns. Further research is needed to guide the clinical approach to patients harboring these patterns. Studies exploring not only the overall impact of therapeutic trials, but targeting the identification of which patients benefit the most from pattern suppression and at what cost (non-sedating drugs versus sedation) are warranted. There are currently ongoing studies investigating the role of pattern suppression with anti-seizure medications, which will soon shed light in the therapeutic management of patients harboring IIC patterns.