Airway management

As with any critically ill patient, the initial approach to the patient in SE involves the basics of airway, respiration, and circulation. Often, the best way to obtain control of the airway is to terminate SE pharmacologically; this should proceed concurrently with the elements of life support. Many patients will need intubation for airway protection and to maintain ventilation during SE treatment. The choice of techniques and medications for intubation depends on several factors; in SE patients, the most important ones include the possibilities of increased intracranial pressure (ICP) and of neuromuscular dysfunction.

The act of laryngoscopy and intubation raises the ICP, which is especially problematic in patients with preexisting intracranial lesions. Premedication with intravenous (IV) lidocaine or thiopental, or probably with etomidate, blunts the rise in ICP and thereby lessens the risks of intubation (4, 25). All three of these drugs probably lower cerebral oxygen demand transiently, resulting in an appropriate decrease in cerebral blood flow; thus, ICP falls without producing an oxygen deficit. Thiopental (1.5mg/kg) and etomidate (0.3mg/kg) are also useful, if only briefly effective, anticonvulsants, and will often cause convulsive activity to cease for several minutes; this facilitates intubation. Lidocaine (1mg/kg) has been suggested as an anticonvulsant as well (11), but in my experience it has not been effective for this purpose. However, lidocaine and etomidate are useful agents to consider in the patient with hypotension or marginal cardiovascular reserve, as they are less likely to cause hypotension than thiopental. Topical or regional anesthesia of the larynx with lidocaine does not confer the same ICP control advantage as an IV dose. A single dose of etomidate is well tolerated, but because of its adrenal suppressant effects it should not be used as a continuous infusion.

Whether to employ neuromuscular junction (NMJ) blockade in a patient with SE remains the subject of debate. Those favoring NMJ blockade suggest that complete relaxation increases the likelihood of successful intubation on the first attempt. The counter argument suggests that NMJ blockade allows SE to continue or recur while the patient is paralyzed (and thus unable to manifest seizure activity, except via the autonomic responses of pupillary dilation and tachycardia), and also leaves the patient in whom intubation is unsuccessful without spontaneous respiratory capabilities for the duration of the blockade. It is my preference to use NMJ blockade. In any case, once one is committed to intubating the patient, appropriate personnel and equipment should be available to accomplish this maneuver expeditiously, to maintain the patient's airway by proper head and jaw positioning, and to provide ventilation via a bag-valve-mask system, regardless of whether NMJ blockade is employed.

The choice of the NMJ blocking agent is also contentious. Some prefer to use succinylcholine (1.5mg/kg), since it provides the shortest latency to intubating conditions (60–90 seconds) and the briefest recovery time (5–8 minutes) of the currently available agents. However, as a depolarizing agent, it carries the risk of profound, sometimes fatal hyperkalemia in patients who have dysfunction of the neuromuscular system, including spinal cord injuries and sometimes hemiparesis of cerebral origin. Thus, unless the patient was known to be neurologically intact until the present episode of SE, a different drug should be considered.

With the withdrawal of rapacuronium from the market, the major choices for a nondepolarizing NMJ blocking agent include vecuronium, rocuronium, and cisatracurium. Vecuronium (0.1mg/kg), the least expensive member of this class, has a somewhat longer latency than succinylcholine and lasts about 20–30 minutes. During this time physical findings in the patient cannot be used as evidence regarding the presence of seizure activity; thus, unless electroencephalographic (EEG) monitoring is in progress, one must assume that the patient is continuing to seize despite the lack of motor activity. Although these drugs are usually considered to be cleared quickly in patients with normal renal function, there is considerable variability in the normal population in the rate of excretion (10). High-dose rocuronium (0.6mg/kg) has a latency similar to that of succinylcholine, but its duration of blockade at this dose is about 30–60 minutes. In contrast to the other NMJ blocking agents discussed here, cisatracurium (0.2mg/kg, latency 2–4 minutes, duration 30–60 minutes) does not depend on renal (or hepatic) clearance mechanisms. However, it does have a potentially epileptogenic metabolite (laudanosine), which is probably of minimal import in humans (20).

The end of NMJ blockade is typically apparent from the return of spontaneous respiratory efforts and other muscle movements, unless they have been suppressed by the medications being used to terminate SE. One can also use a train-of-four nerve stimulator to determine the extent of residual blockade. To speed recovery from these drugs, consider using neostigmine (50–70µg/kg), usually given once two twitches are present on train-of-four stimulation. This will permit examination of the patient sooner; however, its duration of action is shorter than that at some of the NMJ blockers, so caution should be used when extubating patients whose blocking agents were reversed with neostigmine, as its effect may wane before the patient is strong enough for airway protection and breathing.

The choice of a route for intubation is similarly contentious. I favor orotracheal intubation under direct laryngoscopy, even though it often involves NMJ blockade. Blind nasal intubation has been proffered as an alternative that may not require NMJ blocking agents. The risks of sinusitis, and secondarily nosocomial pneumonia, as a consequence of nasal intubation argue against its use (33).