Only 90 years ago, in 1873, Obersteiner, whom in his time knew a lot about epilepsy, used to say that “on a condition which manifests itself so strongly, treatment cannot be effective.” In fact, a useful therapeutic treatment was introduced only about 75 years ago, when Wildermuth (1889) recommended the use of amyl hydrate to treat status epilepticus. Until then, doctors used different methods, which have now all been abandoned, such as cupping glasses, bloodletting, ice bags, quinine (Delasiauve, 1855), ammonia inhalation, potassium bromide (Bourneville, 1869), amyl nitrate (Crichton, 1873) or even opiates, which are unfortunately still used nowadays, despite their complete absence of anticonvulsive effect.
However, if someone attempts to get a clear understanding and reads through the extensive literature on the management of status epilepticus published since, he will likely find it extremely confusing: Indeed, almost all medication that is thought to have a sedative, hypnotic, narcotic, or anticonvulsive effect is being recommended, and this in every possible method of application. Each author has his own method and declares it to be the best, following brief experimentation and applied only to a few cases. There is no comparative analysis of the different methods to be found in all of the literature.
There can be several reasons for this:
- Status epilepticus is in fact a condition that occurs relatively rarely, except maybe at the Sunland Training Center in Florida, where Carter and Merritt (1952) claim to have treated 1,000 cases of status epilepticus over the last 10 years. I doubt anyone would have enough time to gather such a large number of cases for comparison.
- Because of the circumstances during status epilepticus, it is difficult to assess the effect of each treatment. The patient's critical condition, which could lead to death, requires the use of several therapeutic treatments, either simultaneously or one after the other.
- Among the therapeutic treatment rules committed to manuals, many are not actually resulting from systematic research but more so from discoveries due to chance, that is when they are not based only on traditions that have been taken for granted and passed on without any kind of evidence.
We therefore begin by providing an overview of the commonly used methods, and then talk about our own experience.
Most Common Medical Treatment and Therapeutic Methods
Alcohol and aldehydes
In this category, one can find products that have been used for a long time for the management of status epilepticus such as amyl hydrate, chloral hydrate, paraldehyde, and more recently, avertine (Rectanol). Amyl hydrate (tertiary amyl alcohol [(CH3)2 C9OH2H5], which was used in around 1900 (Ackermann, 1897; Flügge, 1898; Maab, 1900 quoted in Janz, 1967; Alt, 1905; Böckelmann, 1906) and recommended with a dose of 5–8 g (Wildermuth, 1889), was quickly abandoned and replaced by chloral hydrate [CCl3 CH(OH)2]. This medication, which is still often used although it is now combined most of the time with barbiturates (Chavany et al., 1954; Guiot, 1954; Livingston, 1954; Zappoli, 1956; Dreyer, 1960), is administered intravenously with a 10% solution and a dose of 0.05 g/kg, that is, 3–5 g. Subcutaneous injection is not recommended because it creates local irritations (Poulsson, 1940). When administered rectally, the product becomes progressively effective after 15–30 min and acts over a 2–4 h period (Janz & Kautz, 1963). This medication should not be used in patients with heart failure or high blood pressure, since a high dose could lead to a vasomotor paralysis.
Paraldehyde was recently recommended again by Whitty and Taylor (1949); however, there is a very small margin between the anticonvulsive dose and the lethal dose, and its effect seems short-lived (Zappoli, 1956) and the drug should therefore be administered repeatedly. Intramuscular injection is efficacious only after 30 min to 1 h (Whitty & Taylor, 1949), whereas the effect is instant when injected intravenously. Rectal administration has been recommended with a dose of 4–8 g (McNaughton, 1954; Prokop & Haid, 1959; Kramer, 1960; Coller [date unknown] quoted in Janz, 1967), Livingston believes that this application method is no less efficacious than parenteral administration, which is recommended by all other authors (Wechsler, 1940; de Elio et al., 1949; Williams, 1950; Carter & Merritt, 1952; Elste, 1955; Dreyer, 1960). Parenteral injection is undertaken with a dose of 0.2 mg/kg; it may be diluted in a solvent to facilitate absorption for intramuscular injection (Whitty & Taylor, 1949). In emergency cases, lower doses can be repeated every 30 min. Although paraldehyde is mostly recommended by English-speaking authors as the preferred initial medication, avertine (or Rectanol), a base narcotic often used in surgery, which is a tribromo-ethanol (CBr3 CH2 OH), is recommended as last resort, after all other medication has proven to be ineffective (Gallinek, 1921; Blume, 1929; Schoene, 1937 quoted in Janz, 1967; Alpers, 1946; Roger et al., 1950; Riser, 1952; Hentschel, 1956; Gastaut, 1963).
Among all medications used in the management of status epilepticus, phenobarbital, via parenteral injection, is still in first place (in a soluble salt sodium of phenylethylbarbituric acid).
It is at this stage, from what I know, used almost exclusively as the first medication at home, even though the dose is still often very low. If only practitioners would follow Guillaume and Mazars' recommendation (1956) and administer an intramuscular or intravenous injection of phenobarbital as early as possible at a dose of 0.30 g, we would not, in most cases, need to administer any other injection right after the patient has been hospitalized or need to inject doses as high as 0.60–1 g over the next 6 h. Most of the authors (Grinker, 1920; Ayala, 1926; Satta, 1931; Bridge, 1949; McNaughton, 1954) advise—as did Guillaume and Mazars—starting as early as possible with a high initial dose so that further medication does not have to be fragmented into lower and repeated doses (0.20 g every 30–60 min). In fact, because of the risk of accumulation, we cannot, especially when renal function is impaired, continue to administer high doses for several days. The effect of phenobarbitone (Gardenal) is slow and starts within 10–15 min (Bamberger & Matthes, 1959); however, it does last for a relatively long time.
The use of intraspinal injection inaugurated by Von Gmelin (1923) and later by Italian authors (Grinker, 1920; Ayala, 1926; Patterson et al., 1926; Fumarola and Enderle, 1923 quoted in Janz, 1967; De Luca, 1926; Belloni, 1930) does not appear to have persisted.
This is why a whole series of fast-acting barbiturates was then recommended, such as Evipan (Santangelo, 1935; Mizroukhine & Slivko, 1941), Pernocton (Olshausen, 1932), pentothal (Bridge and Tardieu), sodium amytal (McNaughton, 1954; Fraser et al., 1958; Keith, 1963), and Senocal (Livingston, 1954). Their greatest advantage is that they act quickly. However, this is counterbalanced by the need for the injection to be done slowly, which is often difficult during status epilepticus, and the fact that, used at high doses, all of the medications carry a risk of causing respiratory depression. Therefore, their use is recommended only when an anesthesiologist is available to assist.
One medication has the advantage of acting both fast and over a long period of time; it is Somnifene, it contains two barbituric acids: diethyl-barbituric acid (also known as Veronal) and allyl-isopropyl-barbituric acid. From what I know, Somnifene is the second most commonly used medication in hospitals after barbiturates and the first medication for which therapeutic activity has been studied on a large scale. Clemmesen (1926) did not report a single death of the 61 cases treated; a single injection (2.2 g or 4.4 g) was sufficient to stop status epilepticus in 48 cases. Dr. Stubbe-Teglbjaerg reported (Zappoli, 1956) that he treated about 500 cases of status epilepticus with Somnifene over the last 20 years, and that he had only 16 cases of death. The highest dose given over 1 day can be up to 11 g. It is important to note, however, that if treatment is started with an intravenous initial dose of 4 cm3 (Scharfetter, 1929 quoted in Janz, 1967), it should be followed by an injection of 1 cm3 of Somnifene every 30 min (Lieser, 1956) or two injections of 0.10–0.15 g of phenobarbital with an interval of 1 h (Riser, 1925 quoted in Janz, 1967). Dreyer advises to first combine intravenous and intramuscular injection in order to prolong the effect. In any case, Somnifene should be injected slowly; this can be achieved with 500 cm3 of a 40% or 60% dextrose solution (Mezei, 1940; Bamberger & Matthes, 1959).
Following the positive experiences of Murphy and Schwab (1956), it is now easy to find diphenylhydantoin, under the name of Epanutin or Dilantin, in 5 cm3 phials, which contain 250 mg of active substance and which can be used for parenteral injection. Carter (1958) reports a surprisingly high number of cases and results. He claims to have controlled, using a single vial, 97 of 121 cases of status epilepticus in children and 17 additional cases after injecting one more vial. Dreyer (1960), Williams (1950), and Lorentz de Haas (1960) recommended the same dose. Schwab and Murphy (1959) use higher doses in adults. They inject no more than 50 mg (1 cm3) per minute and slowly increase the dose, which can eventually reach 500 mg to 1 g. This way, seizures often stop after doses of 200–300 mg. When the action is insufficient, they add 100–150 mg of phenobarbital. In addition to the slow intravenous injection, they advise administration of 100 mg of diphenylhydantoin via intramuscular injection. This is done to maintain a sufficiently high blood level because of the drug's relatively slow resorption. However, we do not know yet if the results that Carter (1958) reported will withstand critical examination. Carter himself seems to have stopped using diphenylhydantoin. In a later publication Carter (1962) no longer recommends diphenylhydantoin but replaces it with phenobarbital and urea for the management of status epilepticus; he declares at the beginning that diphenylhydantoin is not satisfactory because although it “shortens” somehow the duration of status epilepticus, it is sometimes incapable of stopping an acute case.
Since Thorne reported in 1945 that he managed to stop six cases of status epilepticus with 1 g of tridione either via subcutaneous or intravenous injection, this medication has been recommended despite the lack of critical verification on a larger scale (Erickson et al., 1946; Perstein & Andermann, 1946; Kaufmann & Isenberg, 1952; McNaughton, 1954; Gastaut, 1963). According to later correspondence (1948), up to now, Thorne has managed to rapidly stop 16 of 17 cases using tridione, using up to 7 g of the product within a 3 h time frame. In another case, Perstein and Andermann (1946) gave up to 24 g in a single day without any noticeable drawback.
Dreyer reports, however, that the use of tridione did not have any action in the treatment of grand mal seizure; same for Lorentz de Haas who also consider its action somehow doubtful.
The use of Novocain or its derivatives (Lignocain, Xylocaine, Lidocaine, Paurachiol) in the treatment of status epilepticus was mostly recommended by surgeons (Benvenutti, 1934; Ezes, 1948; Schmahmann, 1948; Laine, 1950; Roger et al., 1950; Lereboullet, 1951; Tardieu & Dalloz, 1954; Bernhard et al., 1955; Taverner & Brain, 1958; Bohm, 1959; Gastaut, 1963) and does not seem to have persisted: the action is so limited in time (Zappoli, 1956) that in the end no result can be obtained without the use of anticonvulsants. It seems to have a more effective action on Jacksonian seizures or in repeated seizures resembling a state of grand mal. In this case, 15–20 cm3 of 1% solution is slowly injected into the vein or into the carotid (Laine, 1950).
No one doubts the existence of cerebral edema, either at the onset or during status epilepticus (Janz, 1960, 1961). Therefore, any medication for which the aim is to dehydrate the brain can be considered as rational. To that effect, it has been often recommended, in addition to anticonvulsive medication, to inject or perfuse hypertonic solutions, especially 40% or 60% glucose, with doses ranging from 40 to 80 cm3.
However, this method does cause some concern; indeed, the fact that the brain receives fluid and glucose might increase its need for energetic substrates, which is already very high during status epilepticus, and as a consequence prolong the latter. Similarly, oxygen intake as recommended by several authors (Niedermeyer, 1959; Böckelmann, 1906), may also be considered a two-edged sword (Puca). As Ruf (1951) was able to demonstrate, experimental status epilepticus (automatic repetition of seizures) lasts as long as there is oxygen intake; the latter supports the transfers of energy (which is multiplied by 20–50 during seizures), whereas a carbon dioxide intake will stop the process.
However, to be completely accurate, the pathologic increase in brain activity during seizures does not only mean that the brain requires more oxygen but also more energetic substrates (glucose). As shown by Gänshirt's experiments (1960), the elimination of metabolites (a function of how the blood is detoxified) also plays an important part. Therefore, any treatment that aims to maintain the circulation and provide oxygen and glucose to the brain can increase the risk of seizures. The experimental data seem to confirm what Lennox (1960) had anticipated when he mentioned that muscular contractions, starvation, and profuse sweating, when creating acidosis and dehydrating the patient, also to a certain extent lead to an auto recovery.
It would seem therefore more rational to look for a treatment that could create acidosis and dehydration without increasing the blood volume. A treatment of this kind could be carried out using sulphonamide diuretics. The most practical, Diamox, was the first to be recommended by the Austrian authors Pateisky and Petsche (1957) and Prokop and Haid (1959) because as well as eliminating water, this agent creates acidosis and favors the intake of carbon dioxide. Pateisky and Ringel (1956) report a particularly impressive case of status epilepticus that was triggered by cerebral edema caused by hanging. The status was refractory and the patient was in deep coma; however with a single injection of 500 mg of Diamox, without using any other anticonvulsive drug the seizures stopped.
In 1962, the same Carter, who 4 years beforehand had reported on his spectacular success with diphenylhydantoin, reported equally surprising results obtained using urea. According to what he reports, he started a perfusion of urea within 5–10 min after the onset of status epilepticus in 57 cases, and in most of those he managed to stop the seizures within 3–5 min, and never >10 min. These results are so extraordinary that one wonders whether what he describes was in fact what is commonly referred to as status epilepticus. Indeed, the results were rather unusual, as much in the percentage of patients with status epilepticus (18%) as in the number of patients who he treated as such (1,000 cases over a 10-year period). Or even when he reports that an injection of urea 5–10 min after the start of status epilepticus was enough to control seizures in less than 10 min!
In rare cases, such as with tetanus, one can be forced to resort to artificial ventilation (via a device and a tracheotomy). This can be accomplished only after complete muscle relaxation, with curare or suxamethonium (Evanson, 1959; Nisbet, 1959; Prokop & Haid, 1959; James & Whitty, 1961). These are extreme measures and can only be implemented with the help of an anesthesiologist, and is done mostly to eliminate external manifestations of status epilepticus so that breathing and blood circulation can be stabilized and so that the patient can later withstand the use of anticonvulsants.
Indeed, electroencephalography (EEG) shows that critical central activity continues after the administration of curare, as demonstrated by James and Whitty (1961) in a case where they used curare in a patient for 36 h under EEG monitoring so that they were able to keep on treating him with anticonvulsants until the critical activity had completely stopped. This method, which has no direct effect on the seizures themselves but only on their peripheral manifestations, should be considered only as a secondary option and only for particular cases such as when anoxia endangers the life of the patient or when a surgical operation becomes urgent (for example for a Caesarean section). This method is also indicated when seizures are caused by an overdose, blood levels are low, and when breathing is compromised and time is critical. Later on, the best time to cease artificial ventilation should be assessed according to the EEG, because otherwise, as was the case for Evanson, there could be a relapse in status epilepticus.
The use of chlorpromazine on its own or as a lytic cocktail is common and was mostly recommended by gynecologists as a treatment for eclampsia. The use of phenothiazine derivatives for the management of status epilepticus is still reported in studies (David et al., 1953; Flügel, 1953; Panse, 1953; Puca quoted in Janz, 1967; Bonnet, 1956; David et al., 1956; Furtado, 1956), although it was not possible to give many statistics and many other authors have criticized and warned against this method. Not only for experimental or theoretical reasons but also based on clinical experiments (Arnold & St. Hift, 1953; Dogan & Vukadinovic, 1954 quoted in Janz, 1967; Bonafede, 1955; Von Ditfurth, 1955; Hentschel, 1956; Dreyer, 1960).
Indeed, phenothiazines cannot have a clinical action that would efficaciously prevent seizures (McLean et al., 1958; Temkoff et al., 1958). Furthermore the experiment shows that refrigeration itself causes an increase in seizures (Fuhrmann et al., 1954). Lafon et al. (1954) did not manage to stop a case of status epilepticus using artificial hibernation and eventually treated with barbiturates. Dreyer outlined that it would be unreasonable to give a drug that has no proven efficaciousness when the life of the patient is as much at risk as during status epilepticus.
The same can be said about opiates, which “contrary to what is commonly believed, have no anticonvulsive action” (Eichholtz, 1957) and which, in addition, are contraindicated in the management of status epilepticus because of their paralyzing action on respiratory functions (Krause & Schum, 1932; Bridge, 1949; Livingston, 1954; McNaughton, 1954). The use of narcotics series of ether or chloroform, still recommended by Bridge as first line therapy, today is completely abandoned because of its lack of evident anticonvulsive action and because it puts the patient at risk for hypostatic pneumonia.
The replacement of a great quantity of cerebrospinal fluid by air was often mentioned in literature. This was first reported by The Lancet in 1907. However, no statistics are available on this. In one case treated by Prokop and Haid (1959), a maximum replacement by air (600 cc) led to only a transitory diminution of the interval between seizures. In any case, this method can only be considered as last resort, and even then, with the greatest caution, since, in case of an unknown cerebral tumor, there is a danger of coning (Lereboullet, 1951). The idea that, with pneumoencephalography, two goals—one therapeutic and the other diagnostic—could be reached at once and that this could be done as one intervention (Guillaume & Mazars, 1956), often did not prove conclusive because it is not possible to treat cerebral edema by filling the ventricles with air.
When we compare our personal experience with the brilliant results reported in the literature, we have to be humble and admit that what can be learned from these studies is what should not be done. If I mention my personal experiences, however, it is to allow us to learn from our mistakes. It should be noted that almost all of our cases of status epilepticus are desperate situations, which have been referred to our clinic because they could not be treated otherwise. This way, we have treated 45 cases of status epilepticus since 1950; this represents an average of four or five cases per year, and this has been done via many different methods. Forty of these cases can be used for a comparative study. Although we lost four patients (10%) despite all therapeutics, it is interesting to note that we managed to stop status epilepticus immediately in only 14 cases. Seizures stopped after the second try in 7 cases and we had to use several repeated methods for 16 of the patients.
When we try to analyze the mistakes we made, we can identify four main errors:
- The initial medication was often used with doses that were too low.
- We omitted in most the administration of a hydrating therapeutic.
- When status epilepticus was prolonged, the treatment was conducted irregularly and doses were too dispersed and too low.
- We could have prevented many relapses if, after the status epilepticus had been controlled, we had continued to give at regular intervals parenteral injections of an anticonvulsant.
Depending on the method used, we obtained the following results:
- With phenobarbital – we managed to stop only 2 of 16 cases of status epilepticus with an initial dose of 0.2 g of phenobarbital. At the time the treatment was started, patients had experienced only six and nine seizures. In three of five cases, status epilepticus ceased after an initial dose of 0.4 g of Gardenal. It should be added that when we started to treat with an initial dose of 0.2 g of Gardenal, a similar second dose did not give any result.
- With diphenylhydantoin – a dose of 0.25 g of diphenylhydantoin was not sufficient to treat status epilepticus either at the start or during treatment (two cases). We managed to stop four cases out of seven with a 0.50 g dose. The fact that in one case seizures persisted despite an initial dose of 0.50 g of diphenylhydantoin, although the patient had experienced only six continuous seizures, shows that this dosage (twice the amount advised by Carter) cannot even be considered as absolutely efficacious even in easy cases.
- With chloral hydrate–we have managed to stop only two of six cases of status epilepticus with a rectal administration of 3–4 g of chloral hydrate. In eight cases, for which other medication were not efficacious, five were controlled with a supplement of chloral hydrate. The combination of phenobarbital and chloral hydrate (used 10 times) was only effective in high doses (4 g of chloral hydrate with 0.4 g of phenobarbital) in three of five cases. We never obtained satisfying results when combining lower doses, that is, 0.3 g of phenobarbital with 3 g of chloral hydrate (six cases).
- Trimethadione – Trimethadione was never used as an initial medication. In six cases we injected 5 cm3 of tridione after trying vainly with 0.4 g of Gardenal, chloral hydrate, and diphenylhydantoin. Trimethadione immediately stopped three cases of status epilepticus, whereas it had no effect on two other cases. We were relatively impressed with the effect of trimethadione in one case where the patient had already experienced 400 seizures at the time of the injection and had not previously reacted to 0.4 g of Gardenal and 3 g of chloral hydrate.
- Pneumoencephalography – it is difficult to judge the true effect of the removal of some cerebrospinal fluid and the pneumoencephalographic monitoring that were used in 11 patients because in five of these cases, three of which were successfully stopped, we were also administering anticonvulsants. In six cases, pneumoencephalography was used as a last resort after trying all other medications: status epilepticus was stopped 3 times and continued 3 times.
Based on our experience, we cannot single out one ideal medication in the management of status epilepticus. However, a few rules should be followed and there are some mistakes to be avoided.
Treatment is divided into four stages:
- Initial treatment at home
- Treatment in hospital
- Later prophylactic treatment to avoid relapse
- General treatment for coma, respiration, heart, and circulation
Initial treatment at home
The best option is parenteral injection of 0.4 g of Gardenal as early as possible (intravenous or intramuscular or even better half/half). To offer another option to practitioners or give them a choice of another medication is not reasonable. Indeed, no one can be expected to commonly keep injection doses of diphenylhydantoin or Somnifene in their emergency kit. In any case, each time, even when status epilepticus stops, the patient should be hospitalized in the nearest specialist clinic; a pediatric, neurologic, neurosurgical, or psychiatric unit.
Treatment in hospital
When status epilepticus cannot be successfully treated at home with phenobarbital, the hospital therapeutic treatment should be started with a half intravenous half intramuscular injection of 0.5 g to 0.75 g of diphenylhydantoin, with an intravenous injection of 500 mg of Diamox or, alternatively, with 2–3 vials (i.e., 4–6 cm3) of Somnifene half intravenous half intramuscular and an intravenous injection of 500 mg of Diamox.
If status epilepticus persists after half an hour, the same medication should be given with a dose of 0.5 g of diphenylhydantoin or two vials of Somnifene, if needed combined with a 4 g enema of chloral hydrate. A fast-acting narcotic such as Evipan (0.4–0.6 g) intravenous or even an enema of Avertine (0.09 g/kg, i.e., 6–10 g) can be administered.
Later prophylactic treatment to avoid a relapse
As prophylactic therapy, one should continue for 24 h after status epilepticus has stopped, to make parenteral injections of either 0.20 g of phenobarbital or 0.25 g of diphenylhydantoin within progressively longer intervals (3–8 h), and this, before the oral antiepileptic dose becomes active. The oral medication should be taken as soon as the patient regains consciousness and should remain combined with a supporting parenteral medication for approximately 12 h.
General treatment for coma, respiration, heart, and circulation
As soon as status epilepticus starts, the bladder and intestine should be artificially emptied and later followed with regular catheterization of the urinary bladder. After status epilepticus has stopped, dehydration can be treated with perfusions of isotonic or plasma solutions.
The heart and circulation can be treated with ouabaine and lobeline, whereas Cardiazol and Camphor are contraindicated because of their convulsive effect. If the patient has respiratory or circulatory insufficiency, as a very last resort for desperate cases one could revert to artificial ventilation with a tracheotomy. It should then be combined with complete curarization under EEG monitoring.
Antibiotic treatment is always indicated when the length of a prolonged coma represents a risk of hypostatic pneumonia. If the coma lasts for more than 4–5 h after the end of convulsive status epilepticus, there is a high chance of cerebral edema caused by an expanding process or of intracerebral hemorrhage.
However, all of this can only be a guide, which should on the one hand leave sufficient room for individual modifications and on the other hand, be simple enough, since all therapeutic treatment protocols should be easy to implement.