CNS
Tissue hypoxia (decreased O2 delivery and increased demand)
Cerebral edema (angiogenic and cytotoxic)
Increased CBF-CMRO2
Increased intracranial pressure
CSF pleocytosis
Hemorrhage
Cerebral venous thrombosis
Cardiovascular
Hypertension followed by hypotension
Tachycardia
Myocardial ischemia
Arrhythmias
Cardiac arrest
Respiratory
Hypopnea/apnea
Hypoventilation
Aspiration
Pulmonary hypertension and edema
Pulmonary embolus
Metabolic
Acidosis (both metabolic and respiratory)
Dehydration
Electrolyte changes (hyponatremia, hyperkalemia)
Hypoglycemia
Hyperthermia
Skeletomuscular
Rhabdomyolysis
Dislocations
Fractures (bilateral humeral head, compression of the first four lumbar bodies)
Renal
Acute tubular necrosis
Gastrointestinal
Hepatic failure
Hematologic
Peripheral leukocytosis
Disseminated intravascular coagulopathy
Patients that develop SE in the setting of acute brain injury may be particularly susceptible to acquire additional brain injury due to the brain being more susceptible to injury, as well as some of the stabilizing pathophysiologic processes being affected, such as loss of autoregulation leading to a delayed increase in regional cerebral blood flow to compensate for the increased metabolic demands of the seizing brain tissue [23].
Another important characteristic of late-phase status is an electromechanical dissociation that occurs and may lead to misinterpretation of the clinical situation: convulsions may decrease or evolve to minor twitching, although electrical cerebral seizure activity continues as NCSE [24]. Table 8.2 presents a scheme of these evolving stages. It is interesting to note that SE is a dynamic state, with different characteristics, depending on when the patient is examined [25]. Thus depending on the time of observation in the ICU, the patient may have obvious grand-mal convulsions or only subtle twitching of the fingers, abdomen or face or nystagmoid jerks of the eyes or no clinical activity, but being in deep coma state. Although the chances in the ICU are that the intensivist will be notified early, because of vital sign monitoring and frequent examinations by the ICU staff, this may not be the case with a patient who was just admitted for ongoing status. This possibility must be kept in mind and all previously seizing patients, who do not regain consciousness soon, should be monitored with a continuous EEG to exclude ongoing NCSE.
Table 8.2
Clinical and EEG seizure correlation during generalized SE
Time | Clinical activity | EEG activity |
|---|---|---|
Onset | Discrete convulsions | Discrete seizure activity |
Continuous convulsions | Merging seizure activity | |
Minutes | Continuous convulsions | Continuous seizure activity |
Minimal twitching in face or distal extremities | Intermittent suppression between bursts of seizure activity | |
Hours | No muscular activity | Periodic epileptiform discharges on a flat background |
Goals of ICU Management and Treatment Options for SE
Management of ICU seizures and SE should include (1) emergent medical management, (2) termination of seizures, (3) prevention of recurrence of seizures, (4) diagnosis and address of the underlying cause of seizures, and (4) prevention and treatment of complications. In a survey of 408 intensivists from the UK, it was shown that following failure of initial management of resistant SE, a benzodiazepine infusion (35%) or anesthetic induction agent (32%) was the preferred second-line treatments. The majority of respondents (57%) gave anesthetic induction agents within 60 min of the start of SE. Thiopentone was administered in 82% of these cases. Clinical assessment was used to monitor the response to the treatment in almost half the cases. However, in more specialized ICUs, such as pediatric or neurological or neurosurgical units the majority of responders used a cerebral function monitor in addition to the clinical examination, emphasizing the greater experience these physicians had [26]. In a more recent survey of international experts in the treatment of SE with a 50% response rate (60/120), there was consensus for using intravenous lorazepam for the emergent (first-line) therapy of SE in children and adults. For urgent (second-line) therapy, the most common agents chosen were phenytoin/fosphenytoin, valproate sodium, and levetiracetam; these choices varied by the patient age in the case scenarios. Physicians who care for adult patients chose cIV therapy for RSE, especially midazolam and propofol, rather than a standard AED sooner than those who care for children; and in children, there is a reluctance to choose propofol. Pentobarbital was chosen later in the therapy for all ages [27].
As a general principle, the time to initiate treatment and to not undertreat early after onset of SE is a more important factor in determining seizure control and outcome than the specific choice of antiepileptic agent [28, 29]. This is illustrated by evidence that early treatment initiation for SE has been shown to be effective [4, 5] and by the fact that while lorazepam was superior to other options in the treatment of SE, the difference was not tremendous (i.e., when compared to phenobarbital [30]), and second agent add-on therapy in the out of hospital setting was ineffective but higher than usual doses of benzodiazepine in the out of hospital setting were effective in controlling SE [28, 29]. A general treatment algorithm for convulsive SE is presented in Table 8.3. The treatment is divided in stages, based on the response to seizure cessation (clinical or electroencephalographic).
Table 8.3
Treatment algorithm for SE
3A Prehospital measures |
IM midazolam 10 mg or, if not available, lorazepam 4 mg IV, intranasal midazolam or clonazepam |
Provide oxygen by nasal cannula or face mask |
If a history of diabetes, check finger-stick blood glucose, if available. Administer 1 amp of DW 50% IV if <60 mg/100 dl |
3B In-hospital measures |
Stage 1: Emergent initial measures |
Preserve airway and oxygenation by oxygen face mask or intubation, as needed |
Establish IV access |
Order EEG to be available during therapy |
Measure finger-stick blood glucose. Administer 1 amp of DW 50% IV if <60 mg/100 dL and 100 mg Thiamine IV |
Send to the Lab: antiepileptic drug(s) blood levels, electrolytes, basic metabolic panel, serum glucose, complete blood count, total and ionized calcium, magnesium, and arterial blood gases |
At the same time with the above: Immediate benzodiazepines—IV lorazepam 0.07–0.1 mg/kg [4 mg IV with an additional 4 mg if no response, if no weight available] or diazepam 0.15–0.25 mg/kg IV. If no IV access, diazepam 20 mg per rectum or midazolam 10 mg IM, buccally or intranasally. Skip this step and go directly to Stage 2, if patient has already received prehospital treatment with benzodiazepines |
Stage 2: Urgent Control |
Give one of the following two agents |
• Phenytoin loading dose 20 mg/kg IV at 50 mg/min or fosphenytoin 20 mg/kg PE (phenytoin equivalents) IV at 150 mg/min or IM or |
• Valproate 25–40 mg/kg IV load at 1.5–3 mg/kg/min |
• If seizures continue after this step maybe give additional phenytoin or fosphenytoin (load additional 5 mg/kg PE to 10 mg/kg PE; goal serum level 20 mg/dl to 25 mg/dl) or valproate (load additional 20 mg/kg IV) |
• Alternatives if allergies or unavailability of the above prevent administration to consider: levetiracetam 30–70 mg/kg IV (500–mg/min) or Phenobarbital 20 mg/kg IV (rate 100 mg/min) |
EEG connected and running |
Stage 3: Refractory SE |
Intubation and mechanical ventilation |
Hemodynamic support by pressors and IV fluid boluses |
AEDs for RSE: |
• Midazolam 0.1–0.2 mg/kg IV bolus, which can be repeated every 5 min up to total 2 mg/kg, followed by infusion 0.1–2.0 mg/kg/h, suppress seizure activity for 24 h then start weaning |
• If for any reason unable to give midazolam, give Propofol 2 mg/kg IV bolus and 150 μg/kg/min to 200 μg/kg/min infusion, suppress seizure activity for 24 h then start weaning |
• If not used yet and in patients with DNR status may use valproate (dosing as above) |
• Additional AED alternatives include levetiracetam, phenobarbital, or lacosamide |
Need continuous EEG monitoring at this point. NCSE should be managed the same as CSE. |
Stage 4: Super-refractory SE |
Stage 4.1 |
AEDs for (consider one of the following) |
• Pentobarbital 10 mg/kg IV load at up to 50 mg/min, can be repeated several times until EEG burst-suppression pattern with 20–30 s suppression goal is achieved. Start at the same time continuous infusion 1 mg/kg/h and titrate up to 10 mg/kg/h for same goal. |
• Thiopental 2–3 mg/kg IV bolus and 0.3 mg/kg/min to 0.4 mg/kg/min infusion or |
• Additional agents to consider especially if breakthrough or withdrawal seizures are recorded: |
o Ketamine 0.5–4.5 mg/kg bolus IV and up to 5 mg/kg/hour infusion for 24–48 h |
o Phenobarbital 15–20 mg/kg load |
Stage 4.2 |
Can be used after 4.1 measures fail or in parallel with them (in order from the first to the last resort) |
• Isoflurane or desflurane or gabapentin or levetiracetam (in acute intermittent porphyria) |
• Topiramate 300–1600 mg/day per orogastric tube (if no increased stomach residuals) |
• Magnesium 4 g bolus IV and 2–6 g/h infusion (keep serum levels <6 mEq/L) |
• Pyridoxine 100–600 mg/day IV or via orogastric tube |
• Methyl-prednisolone 1 g/day IV for 5 days, followed by prednisone 1 mg/kg/day for 1 week |
• IVIG 0.4 g/kg/day IV for 5 days |
• Plasmapheresis for 5 sessions |
• Hypothermia 33–35°C for 24–48 h and rewarming by 0.1–0.2°C/h |
• Ketogenic diet 4:1 |
• Neurosurgical resection of epileptogenic focus if any |
• Electroconvulsive therapy |
• Vagal nerve stimulation or deep brain stimulation or transcranial magnetic stimulation |
Stage 4.3 |
If several weaning attempts have failed over a period of months, consider end-of-life discussion with family or surrogate decision maker and withdrawal of life support with subsequent autopsy (if no etiology has been found ) |
Emergent Medical Management
Prehospital Management
Management of SE must begin by the emergency medical services in a prehospital setting. Several studies have attempted to assess the possibility of aborting SE even prior to the hospital. In a randomized, double-blinded study, lorazepam was 4.8 and diazepam 2.3 times more effective than placebo in terminating SE on the arrival at the emergency department when given intravenously (IV) by paramedics [4]. In another prehospital study, midazolam at doses of 2 mg/kg for children or 10 mg for adults intranasally or intramuscularly (IM) was comparable or better than IV diazepam [31]. The RAMPART study was a double-blind, randomized, noninferiority trial comparing the efficacy of intramuscular (IM) midazolam (10 mg followed by placebo IV, n = 448) with that of IM placebo followed by intravenous lorazepam (4 mg, n = 445) for children and adults in SE treated by paramedics. At the time of arrival in the emergency department, seizures had ceased without rescue therapy in 73.4% and 63.4%, respectively, favoring midazolam by an absolute 10% (95% CI 4.0–16.1, p < 0.001). Based on this data the authors concluded that midazolam IM is at least as safe and effective as lorazepam IV for prehospital seizure cessation [5].
If there is a history of diabetes mellitus or insulin use, an emergent glucose measurement should be done by paramedics and if an accucheck is not available, glucose in the form of orally administered juice or sugar or IV injection of DW 50% should be considered. As soon as measurement of glucose becomes available, glucose should be rechecked and thiamine administration also considered.
Oxygen should be administered in a convulsing patient in the ambulance and all prehospital measures to keep an open airway should also be taken.
In-Hospital Management in the Emergency Department or ICU
Management of the convulsing patient who arrived in the Emergency Department or was transferred to an ICU room comprises general medical supportive and diagnostic measures and at the same time specific treatment to terminate the seizures. It is unclear what type of ICU setting is a better fit to deal with SE and current data do not necessarily support admission to a specialized Neuro-ICU. In a single center, non-randomized study of 151 patients (with 168 episodes of SE), 46 (27%) were admitted to an NICU and 122 (73%) to an MICU based on bed availability. APACHE II scores were significant higher in the MICU group (17.5 vs 13.4, p = 0.003) and age in the NICU (58.3 vs 51.5 years, p = 0.041). More continuous EEGs were ordered in the NICU (85 vs 30 %, p < 0.001), where fewer patients were intubated, but more eventually tracheostomized. The NICU had a higher rate of complex partial SE and more alert or somnolent patients, whereas the MICU had a higher rate of generalized SE and more stuporous or comatose patients. Admission diagnoses also differed, with the NICU having higher rates of strokes and the MICU higher rates of toxic metabolic etiologies (39 vs 12% and 11 vs 21 %, p = 0.002). After adjustment, no difference was found in mortality, the ICU or hospital length of stay and modified Rankin score at discharge [32], but discharge status is typically not a great outcome measure as the recovery process is potentially still ongoing and benefits of an intervention or treatment setting may be underappreciated.
General Medical Supportive and Diagnostic Measures for SE
Basic life support with maintenance of airway, breathing and circulation should be provided as soon as the diagnosis of SE is established [6]. Endotracheal intubation (ETI) is important in maintaining adequate oxygenation and preventing aspiration pneumonia. However, few patients with SE require intubation. In the RAMPART study, out of 1023 enrollments, 218 (21%) patients received ETI. Two hundred four (93.6%) of the ETIs were performed in the hospital and 14 (6.4%) in the prehospital setting. Intubated patients were older and men underwent ETI more often than women. Patients with ongoing seizures on ED arrival had a higher rate of ETI, as did those who received rescue anti-seizure medication. Most ETI (62%) occurred early on (prior to or within 30 min after ED arrival). Late ETI was associated with higher mortality (14%) compared to early ETI (3%) [33]. Adequate oxygen supply with a non-rebreather facial mask and airway patency with oral or nasopharyngeal devices may be enough measures for a patient who has had one or more seizures, but has stopped having convulsions. On the other hand, most ICU physicians would intubate a patient in SE for airway protection and for anticipation of administration of respiratory depressant antiepileptics. The goal after intubation is adequate oxygenation (initially 100% FiO2) and ventilation with a goal of normal pH: initial hyperventilation in a paralyzed patient with metabolic acidosis is acceptable, but frequent arterial blood gases are necessary to avoid subsequent respiratory alkalosis, which may further decrease seizure threshold [34].
Paralytics are almost always used for the intubation of a seizing patient, but short-acting agents such as IV rocuronium (0.6–1.2 mg/kg) or vecuronium (0.1 mg/kg) are preferable to succinylcholine, which can induce severe hyperkalemia in neurological patients. As for sedation, thiopental can be used at 3–5 mg/kg, a drug which can also help in seizure control. At least two large IV catheters should be inserted and carefully secured for fluid, drug administration, and withdrawal of blood samples. This is not easy in a convulsing patient: an alternative site such as external jugular catheterization or an alternative route such as intramuscular or rectal administration should be sought [5].
Continuous electrocardiogram, pulse oximetry, and temperature monitoring should be initiated. Non-invasive blood pressure measurements should be started, but the physician should be reluctant to treat elevated pressure during the convulsion phase, unless it is extreme (for example >230 mmHg systolic) or there is suspicion it is the primary cause of the seizures (see the Chapter on hypertension and ICU seizures). Usually, control of the seizures with the first-line AEDs would be enough to reduce the blood pressure. Continuous invasive monitoring of blood pressure should be initiated for all patients in SE as many of the AEDs of choice have strong hypotensive effect, especially barbiturates.
EEG recording should be used to assess the presence or absence of ongoing seizure activity and to direct treatment, but this should not delay initiation of the urgent treatment of SE (see above).
Blood of all patients should be sent for blood glucose, complete blood count, basic metabolic panel, calcium (total and ionized), magnesium, AED levels [35]. On a case-by-case basis work-up should also include a comprehensive urine and blood toxicology panel (focussing on toxins that frequently cause seizures such as isoniazid, tricyclic antidepressants, theophylline, cocaine, sympathomimetics, alcohol, organophosphates, and cyclosporine), liver function tests, serial troponins, coagulation studies, arterial blood gas, and tests to detect inborn errors of metabolism.
For those with diagnosed hypoglycemia (typically via a finger-stick-accucheck), 50 ml of 50% glucose solution should be given. Hyperglycemia may exacerbate neuronal damage caused by SE, therefore glucose should be only administered when lab results confirm hypoglycemia [36, 37]. In case of suspected or confirmed history of alcoholism or other suspected dietary deficiency, 100 mg of IV thiamine is given first along with glucose to avoid precipitating Wernicke’s encephalopathy.
Lumbar puncture is indicated, if there are no signs of increased intracranial pressure or non-communicating hydrocephalus and an infectious process causing SE is suspected. Twenty percent of patients with SE may have “benign postictal pleocytosis” (up to 70 white blood cells/mm3) [38]. This SE-induced pleocytosis is usually polymorphonucleocitic in the differential [20] and this may help differentiate from a primary viral encephalitis as the cause of SE (where the differential is mainly mononucleocytic [7]). This latter case may be resistant to treatment to the point that the term “malignant SE” has been coined [39]. Eight out of 54 (15%) patients with SE may have CSF protein elevation >50 mg/100 ml but in only one case the value exceeded 75 mg/100 ml [20].
Termination of Seizures During SE
Management of an isolated seizure is discussed in another chapter but prolonged or recurring seizures, such as those seen in SE, should be treated aggressively and without delay. The earlier the treatment is initiated, the easier the termination of seizures: 80% of patients had termination of SE when treated within 30 min of onset and <40% when treated after the first 2 h from onset [40]. Compared to the ED, where recurrent convulsive seizures are easily recognized as SE, in the ICU many seizures are nonconvulsive in nature, and are only discovered on EEG. In this situation most intensivists would try to break the convulsion or electrographic seizures by administering benzodiazepines IV (such as lorazepam) for the short-term control of the seizing patient (first-line treatment) and loading the patient with an IV antiepileptic, such as phenytoin or valproate for the long-term control (second-line treatments) (Table 8.3). Because at this point it may not be clear if the flurry of seizures heralds the entry of SE, a very low threshold should exist to intubate the patient in anticipation of more seizures and the need for airway control during the work-up.
If the seizures persist despite 2 AEDs, SE is considered refractory (RSE) and special measures are taken in the ICU (Stage 3, Table 8.3), including the administration of anesthetic doses of short-half-life anti-seizure medications (third-line treatment). In select cases additional dose of second-line AEDs (such as phenobarbital or lacosamide) may be considered but most would start continuous infusions of anesthetic IV drugs , such as midazolam or propofol at this point, which also current guidelines recommend [27, 35]. As mentioned above some question the benefits for anesthetic agents to treat RSE altogether [8], while others consider giving anesthetic agents even earlier in the management course potentially as a second-line agent [35]. If these measures fail, induced barbiturate coma with longer-half-life agents such as pentobarbital is performed which marks the transition to super-refractory (SRSE) (Stage 4). Barbiturate treatments can be used for a short period (24–48 h) with subsequent slow emergence of the patient from coma or, in case seizures recur, for a longer period (usually a week). Additional agents such as ketamine are used at this stage (Table 8.3, Stage 4.1) and additional measures should be considered with various reported success rates based on small case-series at best (Table 8.3, Stage 4.2). Throughout the management and particularly if seizures are difficult to control the underlying cause for SE should be kept in mind and constantly questioned as disease modifying interventions such as plasmapheresis or neurosurgical interventions may be successful in certain scenarios. Prolonged SE may have a very favorable outcome and particularly in young patients with an unclear underlying etiology. Such an example is an 18-year-old man with SRSE who remained in the ICU for 79 days, but subsequently recovered [41]. End-of-life discussions with the family should only be entertained if these extreme and unproven measures fail to control the seizures in the ICU and only after all possible etiologically treatable causes have been considered (Table 8.3, Stage 4.3).
It should be noted that as the intensivist moves down the list of treatment options, the data supporting these treatments become more and more thin and expert-opinion-based. In fact, some of them may be independently associated with worse outcomes. In a recent study, mentioned above, 171 patients were treated over 6 years for SE, of whom 37% were treated with IVADs. Mortality was 18%. Patients with anesthetic drugs had more infections during SE (43% vs 11%; p < 0.0001) and a 2.9-fold relative risk for death (2.88; 95% confidence interval 1.45–5.73), independent of possible confounders (i.e., duration and severity of SE, non-anesthetic third-line antiepileptic drugs, and critical medical conditions) and without significant effect modification by different grades of SE severity and etiologies [8].
The intensity of treatment with therapeutic coma has also been challenged. In another recent study, 50 out of 467 (10.7%) patients with incident SE were managed with therapeutic coma. Therapeutic coma was associated with poorer outcome in the whole cohort: these patients had higher relative risk ratio for new disability (6.86; 95% CI, 2.84–16.56) and higher for mortality (9.10; 95% CI, 3.17–26.16). This effect was more important in patients with complex partial compared with generalized convulsive or nonconvulsive SE in coma. Prevalence of infections was higher (odds ratio, 3.81; 95% CI, 1.66–8.75), and median hospital stay in patients discharged alive was longer (16 days [range, 2–240 days] vs 9 days [range, 1–57 days]; p < 0.001) in subjects managed with therapeutic coma [42].
The intensivist, however, has to remember the more refractory the status is the worst the outcome and these challenging and unproven treatments may be the last resorts to save the life of the patient and stop the seizures. Of the general anesthetics , pentobarbital coma is probably the one treatment with the most challenges (see below in the specific section). However, it may be an efficacious and safe drug to use in super-refractory SE cases. In a recent study from Columbia Presbyterian Hospital in New York, 31 SRSE patients were treated with continuous pentobarbital infusion. Only 8 (26%) of them had a history of epilepsy and 23 (74%) presented with convulsive SE. Underlying etiology was acute symptomatic seizures in 16 (52%, 12/16 with encephalitis), remote seizures in 10 (30%), and unknown in 5 (16%). The mean duration of pentobarbital infusion was 6 days and it controlled seizures in 90% of patients. Seizures recurred in 48% while weaning the infusion, despite the fact that burst-suppression was attained in 90% of patients and persisted >72 h in 56% of them. Weaning was successful after adding phenobarbital in 12 out of 15 (80%) of patients with withdrawal seizures. Complications during or after pentobarbital infusion included pneumonia (32%), hypotension requiring pressors (29%), urinary tract infection (13%), and propylene glycol toxicity and cardiac arrest in one patient each. Interestingly, one third (35%) of patients had no identified new complication after starting the infusion. At one year after discharge, 74% of patients were dead or in a state of unresponsive wakefulness, 16% were severely disabled, and only 3 out of 31 (10%) had no or minimal disability. The authors concluded that pentobarbital-induced coma for SRSE leads to infrequent complications, effectively aborts seizures, and may be successful even after weaning if combined with phenobarbital [43].
All of these studies are observational case series and not randomized controlled trials and therefore the reported associations between treatment and outcome will always remain circumstantial at best. At the heart of it, the chicken-and-egg problem with underlying etiology and SE on the one side and SE treatment on the other side (all being related to each other and affecting the ultimate outcome to different degrees), will remain unresolved in an observational data set. Clearly patients with RSE will not spontaneously stop seizing and recover without intervention. In the following sections we will review the available medication options and the rationale for their use .
Rationale for Using Specific Antiepileptic Medications
Treatment of recurrent seizures and status epilepticus requires fast drug absorption and therefore parenteral administration is essential. Among the currently available standard antiepileptics, only phenytoin, phenobarbital, levetiracetam, lacosamide, and valproate are available in injectable preparations. In addition, benodiazepines such as diazepam, lorazepam and general anesthetics (such as pentobarbital, thiopental, midazolam, and propofol) are available in parenteral forms. In order to act rapidly, the drugs need to cross the blood-brain barrier readily. This is the case with most drugs that are effective in acute seizure management: they are highly lipid-soluble and thus cross in seconds to minutes. High lipid solubility also leads to redistribution from the central compartment (blood and extracellular fluid) to peripheral compartments (fat and organs). The redistribution leads to a drop in plasma concentrations. Therefore repeat infusions are necessary to maintain adequate plasma levels. Continuous administration increases the concentration of the drug in the central compartment and leads to saturation of the peripheral compartment to the degree that the drug no longer redistributes. If drug administration ceases, plasma levels will be maintained by diffusion from the peripheral to the central compartment, which may result in unfavorable side effects, such as prolonged obtundation or cardiorespiratory collapse. These effects are dangerous and account for some of the morbidity and mortality associated with SE [44]. When administered in an ICU setting with readily available central access for urgent administration of IV fluids and vasopressors to counteract any drug-induced hypotension, this side effect is not associated with worse outcome [45].
The rationale for using benzodiazepines as first-line drug was based on a number of randomized controlled trials. The first randomized, double-blind study was conducted by Leppik et al., who compared diazepam to lorazepam in patients with SE. Both drugs were highly efficacious at controlling the seizures (see below) [46]. Another randomized, non-blinded clinical trial compared a combination of diazepam and phenytoin to phenobarbital in 36 patients with generalized convulsive SE. The cumulative convulsion time had a strong trend to be shorter for the phenobarbital group than for the diazepam/phenytoin group (median 5 vs 9 min, P < 0.06). The response latency (elapsed time from the initiation of therapy to the end of the last convulsion) had also a tendency for being shorter for the phenobarbital group (median 5.5 vs 15 min, P < 0.10). The frequencies of intubation, hypotension, and arrhythmias were similar in the two groups [47]. The results of his study, although not reaching statistical significance due to the small number of patients, provided evidence of the safety and efficacy of phenobarbital, but did not convince the majority of the medical community , who preferred shorter acting agents with a safer clinical profile.
Ten years later, the landmark study from the Veterans Affairs Status Epilepticus Cooperative Study Group was published [30]. It was a randomized, double-blind, multicenter trial from 16 VA medical centers of four IV regimens, either for overt SE or subtle SE: diazepam (0.15 mg/kg) followed by phenytoin (18 mg/kg), lorazepam (0.1 mg/kg), phenobarbital (15 mg/kg), and phenytoin alone (18 mg/kg). Interestingly, lorazepam followed by phenytoin, the most commonly used combination today, was not included. If the first treatment had failed, an algorithm to follow with a second and third treatment regimen was also available. Treatment was considered successful when all motor and EEG seizure activity ceased within 20 min after the beginning of the drug infusion and when there was no return of seizure activity during the following 40 min. Five hundred seventy patients were enrolled. Three hundred eighty-four patients had verified overt convulsive SE and 134 subtle SE. An important finding was that SE has to be controlled with the first antiepileptic agent: the second agent was successful in only 7% of cases when the first one failed. In the convulsive SE group, lorazepam was successful in 64.9% of patients, phenobarbital in 58.2%, diazepam plus phenytoin in 55.8%, and phenytoin in 43.6% (P = 0.02, but in the intention-to-treat analysis only with a trend). Lorazepam was significantly superior to phenytoin in a pair-wise comparison (P = 0.002). In the subtle SE group no significant differences among the treatments were detected (17.9%, 24.2%, 8.3%, and 7.7%, respectively, for the four regimens, P = 0.18, in the intention-to-treat analysis P = 0.91). There were no differences among the treatment groups with respect to recurrence during the 12-h study period, the incidence of adverse reactions (hypoventilation, hypotension, cardiac arrhythmias), or the outcome at 30 days. However, comparing the two types of SE, outcomes for subtle SE were significantly worse at 30 days (50.1% of patients with overt SE were discharged from the hospitals vs only 8.8% of those with subtle SE, P < 0.001). Similarly, hypotension requiring treatment occurred more often in patients with subtle SE (p < 0.001). During the first 12 h after the end of the infusions, no patient with subtle SE regained consciousness, compared to 17% of patients with overt SE (but with no significant difference among the four treatment groups).
At 30 days, the outcome of patients who responded to the first-line drug in both the overt and subtle SE groups was better than those who did not respond in the Veterans Affairs Status Epilepticus Cooperative Study Group [30]. Mortality in the non-responders was twice as high as that in the responders. Based on these results, Treiman and colleagues concluded that lorazepam was more efficacious than phenytoin in overt SE treatment and overall easier to use than the other regimens. Also based on these results, various treatment algorithms have been proposed which combine treatment first with lorazepam and then with phenytoin within the first 30 min after SE onset [48].
Recently a small French study compared a single to an early double antiepileptic agent approach for the treatment of SE in the out of hospital setting (68 patients in each arm). This randomized, double-blind, placebo-controlled trial explored the efficacy of levetiracetam as add-on therapy to benzodiazepines (clonazepam 1 mg) for generalized CSE [29]. This study found no difference for the primary outcome measure defined as cessation of convulsions at 15 min after drug administration and also no difference in the predefined post-hoc analyses of secondary safety and efficacy endpoints. Impressively, 84% of patients in the control arm who only received 1 mg of clonazepam were seizure free , which is higher than the 59% and 65% for intravenous lorazepam (4 mg or 0.1 mg/kg) and 73% for intramuscular injection of midazolam (10 mg) [4, 5, 30]. Patients were allowed to get an extra dose of clonazepam and possibly the higher dosed benzodiazepine should be credited with the high response rate.
After these measures fail and if the patient is in the ICU, anesthesia with midazolam or propofol is suggested for treatment of RSE. Alternatively, phenobarbital is tried first for the next 30 min, before one proceeds to general anesthesia. The notion, however, is to individualize the treatment to the patient, than follow a strict, inflexible algorithm: for example, there are selected patients with good response to IV lorazepam, who may benefit from subsequent oral administration of the drug instead of an additional medication [49]. When the first-line drugs fail to control SE, the subsequent choices have markedly reduced efficacy [30], either due to intrinsic refractoriness or delay of treatment with reduced probability for response [50]. Until now we do not have a way to predict which patient will not respond to treatment and for whom the intensivist should, for example, skip treatment steps and go directly to general anesthesia. Mayer et al. examined the issue of predictive factors for refractory SE in a retrospective study of 74 patients with 83 episodes of SE. Refractory SE was defined as seizures occurring >60 min despite treatment with benzodiazepines and an adequate loading of a second standard IV antiepileptic. In 57 (69%) episodes seizures occurred after benzodiazepine treatment and in 26 (31%) even after a second agent was administered (i.e., fulfilling the criteria for RSE). NCSE and focal motor seizures at onset were independent risk factors for RSE in the multivariate analysis (odds ratio 11.6, 95% CI 1.3–11.1, P = 0.03 and 3.1, 1.1–9.1, P = 0.04, respectively) [51].
However, there is no standardized management of RSE even among neurologists specializing in critical care. A survey among 63 (out of 91 participants who responded) experts in this field from Austria, Germany and Switzerland found that two thirds would apply another non-anesthetizing drug (such as phenobarbital) for both convulsive and complex partial SE after the failure of first-line drugs. A general anesthetic was more often used in convulsive than in complex partial SE as an alternative (35% vs 16%, P = 0.02). All participants would proceed to general anesthesia for ongoing seizures after these measures had failed, in case of CSE and, interestingly, 75% of them in case of NCSE. One third of participants would not use EEG, but only aim for clinical seizure termination. The vast majority (72%) responded that they would start weaning general anesthesia within 24–48 h [52]. The more recent international survey of critical care experts, however, suggested that for adult patients the preference was to choose continuous IV therapy for RSE, especially midazolam and propofol, rather than a standard AED, which may reflect a shift in the treatment paradigm [27].
In the following sections we will present the individual drugs used in the treatment of ICU seizures. Some of the most important data regarding pharmacokinetics, adverse effects, and efficacy, based on published studies pertinent to the ICU , will be presented to the interested reader. Table 8.4 presents an overview of these medications. A more in-depth analysis can be found in standard Epilepsy and Pharmacology textbooks.
IV Loading dose | Maximum Rate | Maintenance (po-IV) | T 1/2 | Elimination | |
|---|---|---|---|---|---|
Diazepam | 0.15–0.25 mg/kg | 5 mg/min | 24–57 h | Hepatic | |
Lorazepam | 0.05–0.1 mg/kg | 2 mg/min | 8–25 h | Hepatic | |
Midazolam | 0.1–0.3 mg/kg | 4 mg/min | 0.08–0.4 mg/kg/h | 1.5–4 h | Hepatic |
Clonazepam | 1 mg (repeat × 4) | 2 mg/min | 10 mg/day | 20–40 h | Hepatic |
Phenytoin | 15–20 mg/kg | 50 mg/min | 4–5 mg/kg/day | 12–48 h | Hepatic |
Fosphenytoin | 15–20 mg PE/kg | 150 mg PE/min | 4–5 mg PE/kg/day | 10–15 min | Hepatic, RBC, tissues |
Lidocaine | 1.5–2 mg/kg | 50 mg/min | 3–4 mg/kg/h | 1.8 h | Hepatic |
Lacosamide | 100 mg | 1–2 mg/min | 50–400 mg/day | 13 h | Renal |
Levetiracetam | 1500 mg (up to 3 g) | Within 15 min | 1–3 g/day | 7 h | Renal |
Valproic acid | 10–25 mg/kg | 1.5–3 mg/kg/min | 15–50 mg/kg/day | 7–18 h | Hepatic |
Thiopental | 2–4 mg/kg | 250 mg/min | 3–5 mg/kg/h | 14–34 h | Hepatic |
Pentobarbital | 6–12 mg/kg | 50 mg/min | 0.5–2 mg/kg/h | 20 h | Hepatic |
Phenobarbital | 1520 mg/kg | 100 mg/min | 1–4 mg/kg/day | 75–120 h | Hepatic, renal (25%) |
Propofol | 1–2 mg/kg | 5 min | 5–10 mg/kg/h initially, reduced to 1–3 mg/kg/h | 0.5–1 h | Hepatic |
Paraldeyde | 5–10 ml rectally | Glass syringes | Repeated in 15–30 min | 3 h | Hepatic, lungs |
Isoflurane | 0.8–2% inhaled | Anesthetic system | Titrate to burst-suppression | Lungs |
Medications Used to Control ICU Seizures and SE (Table 8.4)
Benzodiazepines
Introduction
Benzodiazepines have maintained a significant role as first-line IV treatment for acute seizures or SE since they were shown to be broad spectrum and potent anticonvulsant agents [53]. Their effect is at the synaptic level via the benzodiazepine GABAA receptor complex. They enhance the inhibitory GABA action by increasing the Cl-channel openings and hyperpolarizing the postsynaptic neuron [54, 55]. However, one must keep in mind that first-line anticonvulsants like benzodiazepines and phenytoin fail to terminate convulsive SE in 31–50% of cases [1, 4, 5, 30, 51]. Recent data suggests that possibly the dose of the benzodiazepine as much as the time lapse between seizure onset and medication administration predicts the rate of response, though this was not specifically tested in the trial [29].
Diazepam
Diazepam is a highly lipid-soluble benzodiazepine, which has been used extensively for the treatment of SE, but recently has at least in the USA lost some popularity to lorazepam. It is recommended that diazepam be administered by direct IV injection through a needle or a catheter rather than by infusion. Due to its solubility profile it rapidly enters the brain tissue. However, it redistributes to other parts of the body (fat stores and muscle ) in approximately 15–20 min after it enters the brain. This results in loss of the clinical effect due to a fall in the brain drug levels. Its distribution half-life is 30–60 min and its elimination half-life 24–57 h [56]. Nonetheless, sedative adverse effects are persistent and cumulative in particular with repeated administration, since the drug remains in the fat stores. It has been shown that 5–10 mg/min of diazepam can terminate seizures in 5–10 min in 70–80% of patients. The recommended dose is 10–20 mg (0.15–0.25 mg/kg, at a rate of ≤5 mg/min) [46].
In cases that prolonged IV treatment is recommended for longer term management, the use of an alternative drug is advised. The injectable solution contains 5 mg/ml diazepam in a mixture containing 40% propylene glycol and 10% ethanol and can cause local tissue irritation, venous thrombosis or phlebitis, and pain at the site of injection. Careful monitoring of vital signs is recommended to prevent systemic adverse effects such as hypotension, respiratory depression, profound sedation and coma. Co-administration of other sedatives such as barbiturates can increase the risk of serious systemic side effects [44, 57–59].
Diazepam can also be given by rectal administration. Two controlled clinical studies were conducted to demonstrate the effectiveness of rectal diazepam in treating seizure clusters. The trials were randomized, double-blind, placebo-controlled with the first dose administered at the onset of an identified episode. Seizure frequency was measured over the course of 12 h. Both trials showed that a significantly greater percentage of diazepam-treated patients that ranged from 55 to 62% were seizure free during the observation periods compared with placebo-treated patients. Somnolence was the most commonly reported adverse effect and in over 500 patients treated with rectal diazepam not a single episode of respiratory depression was reported [60, 61]. Despite this favorable drug profile, rectal diazepam administration in the ICU should be considered only in the very few patients without immediate IV access (for example those who, during their convulsions, lose their IV access, continue to seize, and have no obvious veins for cannulation). However, newer benzodiazepines, such as IM midazolam, may be better suited for those problematic administration route cases (see below).
Lorazepam
Lorazepam is closely related to diazepam in terms of efficacy and adverse effects. It has become the drug of choice in the acute management of seizures since the drug is less lipid-soluble than diazepam and subject to less rapid redistribution. Its distribution half-life is <10 min and its elimination half-life 8–25 h [56]. A single injection is highly effective and it has been associated with lower risk of cardiorespiratory depression and hypotension than diazepam. The anticonvulsant effect lasts approximately 6–12 h, making it preferable to diazepam (15–30 min) and particularly appropriate for the management of withdrawal seizures [62]. In a randomized, double-blind trial, lorazepam was compared with diazepam in the treatment of 81 episodes of SE. Patients received one or two doses of 10 mg of diazepam or 4 mg of lorazepam IV. The onset of action did not differ significantly (mean time to end the seizures was 2 min for diazepam and 3 for lorazepam). Seizures were controlled in 89% of the episodes treated with lorazepam and 76% treated with diazepam. Adverse effects, such as respiratory depression, occurred in 13% of the lorazepam-treated and in 12% of the diazepam-treated patients [46]. This slightly superior clinical profile of the drug was also confirmed in the pediatric population. The two drugs were compared in 102 children in a prospective, open, “odd and even dates” trial. Convulsions were controlled in 76% of patients treated with a single dose of lorazepam and in 51% of those treated with a single dose of diazepam. In this study, some patients received lorazepam rectally with 100% efficacy. Significantly fewer patients treated with lorazepam required additional anticonvulsants to terminate the seizures. Respiratory depression occurred in 3% of lorazepam-treated patients and 15% of diazepam-treated patients. Interestingly, no patient who received lorazepam required admission to an ICU [63].
In another retrospective study, efficacy, safety, and cost of lorazepam treatment in 90 episodes of SE was compared to diazepam. Fewer seizure recurrences followed lorazepam administration (given either as first, second, or third dose of benzodiazepine, P = 0.0006). There was no difference in adverse effects or cost. The authors recommended that lorazepam be the first-line therapy in preference to diazepam in adults with convulsive SE [64].
Most importantly, as discussed above, lorazepam was found to be the most efficacious and safe treatment for treatment of SE when compared to diazepam with phenytoin, phenytoin alone, and phenobarbital [30], and also shown to be safely administered in the out of hospital setting [4].
Due to the strong tendency for tolerance following lorazepam treatment, longer-term maintenance antiepileptic drugs must be given in addition . The recommended dose of lorazepam is 0.05–0.1 mg/kg (usually 4 mg), repeated after 10 min if necessary. The rate of injection should not exceed 2 mg/min.
Midazolam
Midazolam is a unique water-soluble compound, whose benzepine ring closes when in contact with serum and converts it into a highly lipophilic structure, crossing rapidly the blood-brain barrier. Its water solubility leads to rapid absorption by intramuscular injection or by intranasal or buccal administration. Midazolam is 96% protein-bound and is metabolized in the liver before renal excretion. It has an ultra-short distribution half-life of <5 min and a short elimination half-life of 1.5–4 h [56]. Thus, its action is very short and seizures may recur few minutes after they have stopped. However, in the ICU the volume of distribution may be expanded and the half-life may be prolonged, especially with liver dysfunction [65]. Acidosis can also reduce the lipid solubility of the drug by opening the benzepine ring structure and thus, decrease CNS entrance and seizure control. Despite these deficiencies, midazolam is probably the best benzodiazepine that can be used as a continuous infusion, because of its favorable kinetics and the lack of propylene glycol as a vehicle (which can cause cardiac arrhythmias). An IV bolus of 0.1–0.3 mg/kg at a rate not to exceed 4 mg/min can be repeated once after 15 min. The recommended rate for IV infusion is 0.08–2 mg/kg/h [35]. Higher rates may be associated with hypotension but when given in a critical care setting rates up to 2.9 mg/kg/h are likely safe [45]. The high water solubility of midazolam and rapid absorption makes it a better agent for IM injection than the other benzodiazepines, when IV administration route becomes a problem in the ICU [66]. The mean half-life of IM midazolam (2 h) is slightly longer than the IV route. IM diazepam and lorazepam have a relatively slower absorption, induce local discomfort or can precipitate at the injection site and are not recommended for the treatment of SE [62]. However, IM midazolam has been successfully used to stop frequent seizures or SE within 5–10 min in children and adults [67–71]. In a prospective, randomized study in the emergency department IM midazolam was compared to IV diazepam in their ability to stop seizures. Eleven patients received diazepam (0.3 mg/kg, maximum 10 mg) and 13 midazolam (0.2 mg/kg, maximum 7 mg). Midazolam was administered faster, because of no need for starting an IV line (mean time from arrival to administration of the drug was 3.3 vs 7.8 min, P = 0.001) and resulted in faster cessation of seizures (mean time from arrival to cessation 7.8 vs 11.2 min, P = 0.047) [69]. The usual IM dose of midazolam is 5–10 mg (0.2 mg/kg).
More recently, Ulvi et al. prospectively evaluated midazolam infusion in 19 patients with refractory SE (not responding to initial IV administration of 0.3 mg/kg diazepam (three times at 5-min intervals), 20 mg/kg phenytoin, and 20 mg/kg phenobarbital. These patients were given an IV bolus of midazolam (200 mcg/kg) followed by a continuous infusion at 1 mcg/kg/min. The dose was increased by 1 mcg/kg/min every 15 min until seizures were controlled. In 18 (94.7%) patients, seizures were completely controlled in a mean time of 45 min, at a mean infusion rate of 8 mcg/kg/min. No significant changes in blood pressure, heart rate, oxygen saturation, or respiratory status were noticed. The mean time to full consciousness after stopping the infusion was 1.6 h and the mean infusion duration of midazolam was 14.5 h [72].
A recent randomized controlled trial demonstrated that intramuscular injections of midazolam as first-line treatments for convulsive SE were in the out of hospital setting at least as effective and safe as intravenous administration of lorazepam [5]. One of the advantages of this approach is that midazolam unlike lorazepam does not have to be refrigerated making it a more practical choice to use in the field for many emergency services.
Phenytoin and Fosphenytoin
Phenytoin is insoluble in water, and the parenteral formulation contains 40% propylene glycol, 10% alcohol, as well as sodium hydroxide to adjust the pH to 12. This solution is highly caustic to veins and it may cause necrosis to the surrounding tissues by extravasation. The rate of administration has been limited to a maximum of 50 mg/min, although in clinical practice it is given more slowly—over 25–45 min in the adult patient—to minimize the pain at the injection site and reduce the risk of cardiovascular toxicity from the propylene glycol diluent. It should be mixed only with normal saline and other drug administration through the same line should be avoided. As a lipid-soluble compound, phenytoin readily enters the brain (it reaches peak levels within 15 min) and its redistribution out of the CNS is slower than the benzodiazepines [73]. The drug is 96% protein bound and competes with other highly bound medications. With low albumin levels, one should consider measuring free instead of total phenytoin levels. Fast infusion of the drug carries the risk for hypotension and QT prolongation, therefore ECG and frequent blood pressure measurements are recommended. Pain, edema, and distal to the infusion site ischemia characterizes the “purple-glove” syndrome, which may occur in 9/152 (5.9%) patients who received phenytoin through a peripheral IV line [74]. There may be a delay of several hours between the infusion and the clinical presentation of the syndrome, which makes the recognition difficult. Nevertheless, phenytoin is a highly effective drug in treating SE [75].
Fosphenytoin sodium is a phosphate ester prodrug of phenytoin that was developed as a replacement for parenteral phenytoin and was approved in the US market in 1996. After administration, phenytoin is cleaved from the prodrug by phosphatases found in the liver, red blood cells, and many other tissues. The conversion rate is not affected by age, hepatic status or the presence of other drugs. Unlike phenytoin, fosphenytoin is freely soluble in aqueous solutions, including IV solutions. It is supplied as a ready-mixed solution of 50 mg/ml in water for injection and is buffered to a pH 8.6–9.0. This relatively lower pH of the vehicle for fosphenytoin is responsible for the lack of local adverse side effects at the injection site as opposed to the highly alkaline IV phenytoin solution. Fosphenytoin can be administered IV or IM and it is extensively bound (~95%) to plasma albumin. The dosage of the drug is expressed in phenytoin equivalents (PE). Seventy-five mg of fosphenytoin results in 50 mg of phenytoin in the serum after the enzymatic conversion ; 75 mg of fosphenytoin is therefore labeled as 50 mg phenytoin equivalent (thus 15 mg PE of fosphenytoin is the same as 15 mg of phenytoin) [76]. The drug is administered IV or IM at doses corresponding to customary phenytoin sodium loading (15–20 mg PE/kg) and consistently produces therapeutic plasma phenytoin concentrations (total 10–20 μg/ml and free 1–2 μg/ml). A maintenance dose of 4–7 mg PE/kg can be given either IV or IM. Therapeutic phenytoin concentrations are attained in most patients within 10 min of rapid IV fosphenytoin infusion (up to 150 mg PE/min) and within 30 min of slower IV infusion (<100 mg/min) or IM injection. Maximal total plasma phenytoin concentration increases with increasing fosphenytoin dose, but is less affected by increasing the infusion rate at a given dose level. It is recommended, following fosphenytoin administration, that phenytoin concentrations not be monitored until complete conversion of fosphenytoin to phenytoin is established. Since the conversion half-life is approximately 10–15 min [77], conversion is completed within 1–1.5 h; serum phenytoin peaks at 30 min following the start of IV fosphenytoin infusion and at 3 h after IM injection.
Fosphenytoin has fewer local adverse side effects (pain, itching, or burning at the site of injection) when given IV or IM compared with IV phenytoin. The most common CNS side effect incidence, such as nystagmus, somnolence, ataxia, and headache does not differ between phenytoin and fosphenytoin [78]. Fosphenytoin has been associated with hypotension in 7.7% of patients, which rarely lead to an intervention and with higher pruritus than phenytoin [79]. Phenytoin administered at fosphenytoin rates can lead to cardiac arrest [80], therefore intensivists have to be very careful while prescribing the drug in the ICU during emergencies. Paresthesias of the lower abdomen, back , head, or neck have been reported with fosphenytoin in particular, when high doses and rapid infusion rate were used. They rapidly resolve without sequelae. A possible explanation is the competitive displacement of derived phenytoin from plasma protein binding sites by fosphenytoin. Earlier and higher unbound phenytoin plasma concentrations, and thus an increase in systemic adverse effects, may also occur following IV fosphenytoin loading doses in patients with a decreased ability to bind fosphenytoin and phenytoin (renal or hepatic disease, hypoalbuminemia, the elderly). Close vital sign monitoring and reduction in the infusion rate by 25–50% are recommended for these, frequently encountered, ICU patients [81].
Valproic Acid (VPA)
Valproate is an antiepileptic drug with broad spectrum activity against absence seizures [82], generalized tonic-clonic seizures [83], focal seizures [84, 85] and myoclonic seizures [86].
The drug has enjoyed increasing popularity in the ICU, especially after the introduction of the parenteral formulation. Although VPA is safe and generally well tolerated, there have been early reports of altered hepatic function and of several fatalities in patients taking VPA in combination with other antiepileptics [87]. Careful monitoring of hepatic function is required in patients being treated with VPA, but dose reduction alone may be effective in preventing hepatic complications. In order to provide information on which patients are at risk for VPA-induced hepatotoxicity, Dreifuss et al. conducted a retrospective review of all reports of fatal hepatic dysfunction received by Abbott laboratories between 1978–1984. Patients found to be at the greatest risk for developing fatal hepatotoxicity where children <2 years treated with multiple antiepileptics, and who had other medical conditions, congenital abnormalities, mental retardation, developmental delay, or other neurologic diseases [88]. From 1980 to 1986 the number of VPA-related hepatic fatalities had declined from eight to one, while the number of patients treated had increased nearly six-fold. Nevertheless, VPA is relatively contraindicated in cirrhosis or hepatic failure where it can accumulate and further promote liver damage (see below). Additional side effects include a dose-related thrombocytopenia, platelet dysfunction and coagulopathies (4% incidence in children [89]), pancreatitis, and elevated ammonia.
Valproate sodium injection (Depacon) is approved for use when clinical factors make oral administration difficult or impossible. The pharmacokinetics of the oral and parenteral forms are similar, but if fast therapeutic levels is the goal, like in many ICU situations, the IV form has significant advantages. It can be delivered at a more physiological pH, does not require organic solvents, and has a wider range of solution compatibility compared to phenytoin. In addition, it does not cause sedation or respiratory compromise like the barbiturates or benzodiazepines and has a safer hypotension profile compared to phenytoin and the barbiturates [90]. The drug has an elimination half-life between 7.2 and 17.7 h in studies given to healthy volunteers and this may be due to its 90–95% plasma protein binding [91–93]. Depacon was approved for infusions up to 10–15 mg/kg at 1.5–3 mg/kg/min in the absence of other antiepileptics and in patients valproate-naive. IM injection may produce muscle necrosis and should be avoided.
In a multicenter, open-label trial examining safety of IV VPA, 318 patients with previously treated epilepsy were enrolled; a need for parenteral VPA therapy for various reasons was documented. The median dose of IV VPA was 375 mg given over 1 h. Fifty-four (17%) patients experienced transient adverse effects, such as headache (2.4%), local reactions (2.2%), somnolence and nausea without vomiting (2.2%). The side effects led six patients to withdraw from the study [93]. However, these recommended doses generally result in sub-therapeutic levels of VPA, and they have been challenged in subsequent studies. In a small study by Venkataraman and Wheless, 24 infusions of IV valproate were carried out electively in 21 patients with epilepsy. The dose ranged from 21–28 mg/kg (mean 24.2 mg/kg) and was weight-adjusted. Target infusion rates were 3 or 6 mg/kg/min, i.e. over 4–8 min. No significant BP changes or EKG abnormalities were reported. Based on these results the authors suggested a rate of 6 mg/kg/min [94]. Doses up to 40 mg/kg [95] have been given without serious side effects including significant changes in blood pressure or electrographic abnormalities or respiratory depression. This is in contrast to other commonly used parenteral antiepileptics, such as lorazepam, phenobarbital, diazepam, and fosphenytoin, which have been variably associated with hypoventilation, cardiac arrhythmias, or hypotension [30, 96].
Of special concern in the ICU is co-administration of VPA with phenytoin, two highly albumin-bound AEDs. Two factors, hypoalbuminemia and VPA co-administration with phenytoin have been shown to increase free phenytoin fraction, and a combination of these two markedly increased free PHT fraction. In this case, free levels of both drugs should be measured in addition to total levels [97].
Intravenous VPA has not been approved for the management of status epilepticus (SE). However, the use of IV valproate in SE has been reported in the medical literature in both children and adults [98–102] and in a rat model of SE induced by intra-hippocampal application of 4-aminopyridine [103]. It has been used in both CSE and NCSE.
Price evaluated 24 neurosurgical patients with generalized CSE refractory to diazepam, treated with IV VPA, either as a bolus of 400 mg followed by infusion of 100 mg/h or 15 mg/kg load followed by 6 mg/kg/h infusion. Seizure control was achieved in 6/15 (40%) patients within 2 h in the first group and in 7/9 (78%) patients within 1 h in the second. Only one patient developed thrombocytopenia, without a clear cause-and-effect relationship ever established [104].
In one study conducted in Europe, the efficacy of IV sodium valproate was evaluated in 23 valproate-naïve adult patients with SE (8 with convulsive and 15 with nonconvulsive). A loading dose of 15 mg/kg followed by 1 mg/kg/h infusion led to VPA levels of 68.5 mg/l at 1 h, which was deemed satisfactory. Disappearance of SE in <20 min was considered successful, while in >30 min was considered a failure. Use of IV valproate resulted in the resolution of SE in 19 (83%) patients (7/8 with convulsive and 12/15 with NCSE). All four patients who failed to respond to VPA, as well as to other antiepileptics, had SE secondary to cerebral lesions. There were no relapses of SE within the first 24 h. All patients showed a slight reduction in systolic blood pressure and heart rate, but none required treatment for that. The serum concentrations varied most in four patients older than 80 years, but valproate was still well tolerated. Despite these promising results, the authors suggested that IV valproate be used cautiously in the elderly [101].
Another study assessed the safety and efficacy of IV valproate in 35 patients with SE. Twenty patients had failed treatment with benzodiazepines, and three patients had failed phenytoin treatment. SE was interrupted in 27/35 (77%) patients; the majority of them responded during the bolus infusion. Among the eight patients considered treatment failures, five patients were also refractory to other antiepileptic drugs, two patients responded to an increased valproate dose, and one patient responded to clonazepam. Two patients developed nausea and allergic skin rash after the VPA in these series [105].
In a small retrospective review of hospital records, 13 patients with SE and hypotension received IV VPA therapy. Mean age of patients was 74 and the mean loading dose of VPA was 25 mg/kg (range 14.7–32.7), at a mean rate of 36.6 mg/min (range 6.3–100). There were no significant changes in blood pressure, pulse, or use of vasopressors, suggesting that VPA loading at these high rates is well tolerated, even in patients with cardiovascular instability. Seizures were controlled in 4 (31%) patients with IV VPA, but eventually all patients died as a consequence of their underlying illness (six were post-anoxic and 3 had stroke) [96]. The same group presented their results of using IV VPA in 30 patients on a later occasion. Control of seizures was achieved in 5/11 (45%) of patients with overt convulsive SE, 2/6 (33%) of patients with subtle SE, 4/8 (50%) patients with complex partial SE and all four (100%) patients with simple partial SE. Among patients with overt convulsive SE, the mean duration of SE prior to treatment in patients who responded was 2.6 h vs 36 h in those who did not respond [106].
Based on a review of the available literature until mid-2000, Hodges and Mazur suggested 3 clinical situations, where IV VPA could be considered as a third or fourth-line agent for the treatment of SE (Table 8.5). We also added a use in patients who have a well-documented allergic reaction to phenytoin or fosphenytoin .
Table 8.5
Indications for use of IV valproic acid in SE
1. As adjunctive agent, after benzodiazepines and phenytoin/fosphenytoin have been properly given and while preparations are being made for third-line agents (propofol, midazolam or barbiturates) |
2. Once third-line agents have been given without complete cessation of SE |
3. Instead of third-line agents in patients who do not wish to be mechanically ventilated |
4. Patients allergic to one or more other antiepileptics |
5. Absence SE or myoclonic SE as first or second-line agent |
This sequence of treatment options may be challenged by new data comparing phenytoin to VPA. In a retrospective study of 63 patients treated with IV VPA (average dose 31.5 mg/kg, range 10–78 mg/kg) because of allergy to phenytoin, myoclonus or refractoriness to the other AEDs, Limdi et al. reported 63.3% efficacy, which was increasing with the order that the drug was used (higher as a 4th AED than as a 1st AED). The rate of administration in this study reached 500 mg/min in the majority of patients, with minimal adverse events [107]. In a prospective study from India, Misra et al. evaluated 68 patients with CSE who were randomly assigned to two treatment groups, either VPA 30 mg/kg IV over 15 min or phenytoin 18 mg/kg IV at a rate of 50 mg/min. Interestingly, no benzodiazepines were used before VPA or phenytoin. If seizures failed to be controlled after this first-line treatment, the other agent was subsequently used. Seizures were aborted in 66% in the VPA group and 42% in the PHT group. As a second-line treatment in refractory patients, VPA was effective in 79% and PHT was effective in 25%. The side effects in the two groups did not differ [108].
This study is the first to demonstrate superiority of VPA over phenytoin, but generalizability is questionable as skipping a first-line benzodiazepine clearly deviated from generally accepted standard of care. Further, these results were not replicated in a subsequent randomized study, which compared 50 patients treated with IV VPA with 50 age and sex-matched patients treated with phenytoin, but after administration of benzodiazepines without success. Intravenous VPA was successful in 88% and IV phenytoin in 84% of patients of SE (no significant difference), with a significantly better response in patients of SE <2 h. As in the study by Misra et al., the total number of adverse events did not differ significantly between the two treatment groups [109]. Lastly, in a prospective, quasi-randomized open-label study, Gilad et al. treated patients who presented in the emergency department with SE or acute repetitive seizures with either IV VPA 30 mg/kg or IV phenytoin 18 mg/kg over 20 min in a 2:1 ratio. No benzodiazepines were initially used and in case of failure of the first drug to control seizures, the other one followed. Seventy-four adult patients participated in the study, 49 in the VPA and 25 in the phenytoin arm. In 43 (87.8%) of the VPA patients, the seizures discontinued, and no rescue medication was needed. Similar results were found in the PHT group in which seizures of 22 (88%) patients were well controlled. Side effects were found in 12% of the PHT group, and in none of the VPA group [110]. In another prospective study from Norway, Olsen et al. treated 41 patients with SE or serial seizure attacks with 25 mg/kg of IV VPA loading dose over 30 min, followed by continuous infusion of 100 mg/h for at least 24 h. All patients had initially received diazepam as first-line unsuccessful treatment. In 76% of the cases (31 of 41), seizures stopped and anesthetic agents were not required [111].
Despite the small numbers and conflicting results , these small unblinded studies show better safety profile of VPA compared to phenytoin, at least equivalent efficacy, tolerability of higher infusion rates (up to 6 mg/kg/min) and may begin to challenge the current treatment algorithms. In fact recent European and US guidelines recommend IV VPA (20–40 mg/kg IV, allowing for additional 20 mg/kg if seizures continue) as an alternative to phenytoin second-line treatment for benzodiazepine resistant SE [6, 35]
Levetiracetam (LEV)
Levetiracetam (LEV) was approved as add-on therapy for refractory partial-onset seizures with or without secondarily generalization. The mechanism of action is poorly understood. It binds to the synaptic vesicle protein 2A (which regulates vesicular traffic and neurotransmitter release), inhibits the N-type high voltage-activated Ca++ channel currents, and suppresses the activity of negative allosteric modulators (such as zinc and beta carbolines) in the chloride influx via GABA and glycine-gated channels (therefore restoring chloride influx) [112–114]. It is metabolized by plasma hydrolysis and not through the cytochrome P450 system. LEV and its inactive metabolites are excreted 60–70% renally and the remaining 30% via the fecal route. Renal elimination is proportional to the renal clearance, and the half-life increases in renal insufficiency.
A significant advantage of the drug is that it virtually has no known interaction with the majority of ICU-used drugs, including other antiepileptics. This was explored in a retrospective study conducted in the NICU at the University of Cincinnati by Szaflarski et al. [115]. These authors analyzed the data of 379 critically ill patients and reported that phenytoin used prior to the NICU admission was frequently replaced with LEV monotherapy. Patients treated with LEV monotherapy when compared to other AEDs had lower complication rates and shorter NICU stays. Older patients and patients with brain tumors or strokes were preferentially treated with LEV for prevention and/or management of seizures .
LEV has been used in the treatment of SE because of its IV formulation advantage, although it has never been FDA-approved for this indication. Rossetti et al. retrospectively analyzed 23 patients with SE treated with enteral LEV [116]. The median daily dose of LEV was 2000 mg (range: 750–9000 mg). Ten patients (43%) responded. Initiation of treatment and dosage were significantly different between responders and non-responders: all responders had received LEV within 4 days after the beginning of their SE episode and were administered less than 3000 mg LEV/day. These authors concluded that LEV may be a useful alternative in SE if administered early, even in intubated patients, and that escalating the dosage beyond 3000 mg/day will unlikely provide additional benefit.
In another study from Jena, Germany, Rupprecht et al. compared 8 patients who received LEV as a second-line agent for NCSE with 11 patients treated with conventional IV medications for NCSE [117]. Those patients treated with LEV showed a marked clinical improvement with final cessation of ictal EEG-activity and clinical symptoms of NCSE within 3 days (mean 1.5 days). The response to conventional treatment was similarly effective but there were severe side effects whereas no relevant side effects in the LEV-treated group were noticed. The authors report no significant differences in hospitalization time, time in intensive care unit and outcome between the LEV group and the control group.
LEV has also been used in patients with refractory SE. Patel et al. published a small series of 6 patients with refractory SE (not responding to at least 2 antiepileptic medications), who eventually responded to enteral LEV (dose range 500–3000 mg/day) within 12–96 h [118].
LEV became also available in a parenteral form. Although it is considered bioequivalent to the enteral form and should be administered IV at a similar to the per os dose, it will build serum levels at a much faster rates (500–1500 mg can be given within 15 min), a potential advantage in case of SE. In fact, even higher infusion rates (up to 2500 mg IV over 5 min) were tolerated in healthy adults, with adverse effects (dizziness, 52.8%; somnolence, 33.3%; fatigue, 11.1%; headache, 8.3%) consistent with the established safety profile for the oral formulation [119]. Additionally, it may be considered an attractive alternative in critically ill patients in general, where the enteral administration may lead to delayed or erratic absorption or when swallowing or nasogastric tube placement is not deemed possible. Parenteral LEV has been used for treatment of SE in a study from Basel, Switzerland [120]. In this retrospective study, Ruegg et al. used IV LEV to treat 50 critically ill patients, 24 of which were in SE. These patients in SE received 20 mg/kg IV LEV within 15 min and in 16 (67%) of them (including 4, who received the drug as first-line treatment for simple partial or NCSE) SE ceased. Except for transient thrombocytopenia in 2/50 patients, no other serious or life-threatening side effects were reported by the authors. These results were replicated in the study by Knake et al., who reported their experience with the use of IV LEV for the treatment of 18 episodes of benzodiazepine refractory focal SE in 16 patients, including four patients with secondary generalized SE [121]. SE was controlled in all patients by the given combination of drugs. Additional antiepileptic medications after the IV LEV were necessary in two episodes.
Related posts:
Infection or Inflammation and Critical Care Seizures
Extreme Hypertension, Eclampsia, and Critical Care Seizures
Brain Tumors and Critical Care Seizures
Fulminant Hepatic Failure, Multiorgan Failure and Endocrine Crisis and Critical Care Seizures
Electrolyte Disturbances and Critical Care Seizures
Global Hypoxia-Ischemia and Critical Care Seizures
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
