Antiepileptic Drugs, 5th Edition



Ethosuximide: Chemistry, Biotransformation, Pharmacokinetics, and Drug Interactions

Francesco Pisani MD*

Emilio Perucca MD, PhD, FRCP (Edin)**

Meir Bialer PhD, MBA***

* Associate Professor of Neurology, Department of Neurosciences, Psychiatric and Anaesthesiological Sciences, University of Messina, Messina, Italy

** Professor of Medical Pharmacology, Clinical Pharmacology Unit, Department of Internal Medicine and Therapeutics, University of Pavia; and Consultant Clinical Pharmacologist, Institute of Neurology, C. Mondino Foundation, Pavia, Italy

*** David H. Eisenberg Professor of Pharmacy, Department of Pharmaceutics, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem, Israel

Ethosuximide was discovered at Parke-Davis Laboratories in the early 1950s. After many decades of clinical use (1), it remains a major therapeutic tool for the management of absence seizures.


Ethosuximide, or 2-ethyl-2-methylsuccinimide (C7H11-NO2), a weak acid (negative log of dissociation constant, 9.3), is a white crystalline powder with a molecular weight of 141.17, a melting point of 64° to 65°C, a chloroform/water partition coefficient of 9 at pH 7, and a water solubility of 190 mg/mL.

Its structure (Figure 67.1) includes a five-membered ring with two negatively charged carbonyl oxygen atoms separated by a distance of about 4.5 A. A ring nitrogen is situated between the two groups, and this feature has been suggested to be required for the anticonvulsant effect (2). Because of the occurrence of a chiral carbon at the 2 position of the succinimide ring, ethosuximide exists as a racemic mixture of two separate enantiomers.


Both in animals and in humans, only a relatively small proportion (about 20%) of an administered dose of ethosuximide is excreted unchanged in urine (3, 4, 5, 6). Ethosuximide is eliminated primarily by metabolism, and the most important pathways are summarized in Figure 67.1. In humans, 30% to 60% of the administered dose is recovered in urine as the isomers of 2-(1-hydroxyethyl)-2-methylsuccinimide, and at least 40% of these are excreted as glucuronide conjugates (3,7, 8, 9, 10, 11). The 2- and 3-hydroxy derivatives and 2-carboxymethyl-2-methylsuccinimide represent less important metabolites (3,12,13). There is no evidence that the metabolites of ethosuximide possess significant anticonvulsant activity, except for a report suggesting weak protective effects of 2-(l-hydroxyethyl)-2-methylsuccinimide against pentylenetetrazol-induced seizures in mice (3). In any case, the finding that the unchanged drug is the predominant plasma component in all species including humans argues against an important pharmacodynamic contribution of ethosuximide metabolites.

Studies in rats suggest that ethosuximide metabolism is mediated primarily by cytochrome P450 (CYP) isozymes, with a major contribution from CYP3A and, to a lesser extent, CYP2E and CYP2B/C (10,14,15). In humans, ethosuximide metabolism is stimulated by concomitant administration of rifampicin (16), phenobarbital, phenytoin, and carbamazepine (17,18), all of which are known inducers of cytochrome CYP3A/4.

Although ethosuximide is an enantiomeric drug, potential stereoselectivity in its disposition has been minimally investigated. In one study, the ratio between the two enantiomers measured by chiral gas chromatography in plasma samples from 33 patients receiving long-term treatment was close to unity (19). This finding is consistent with a similar disposition rate for the two enantiomers.



FIGURE 67.1. Main biotransformation pathways of ethosuximide in humans.



Ethosuximide is available clinically only in oral dosage forms. After intake of single doses (two 250-mg capsules), the drug is absorbed relatively rapidly and reaches peak plasma concentrations within 3 to 5 hours in both children and adults (4,20,21). The bioavailability of syrup and capsules is equivalent, but absorption occurs at a faster rate with the syrup (4) (Figure 67.2).

Being a low-clearance drug, ethosuximide does not undergo significant liver first-pass metabolism, and its gastrointestinal absorption is generally assumed to be complete, even though bioavailability studies using an intravenous reference standard in humans have not been reported. Studies in dogs (22) and monkeys (23,24)


found that the absolute oral bioavailability is virtually complete.


FIGURE 67.2. Mean plasma ethosuximide concentrations after single oral doses of ethosuximide (500 mg) as a syrup (solid line) or as capsules (dotted line) in children. (From Buchanan RA, Fernandez L, Kinkel AW. Absorption of elimination of ethosuximide in children. J Clin Pharmacol 1969;7:213-218, with permission.)


The apparent volume of distribution of ethosuximide, calculated by assuming complete oral bioavailability, is approximately 0.7 L/kg in both children and adults (4,17,21), a finding suggesting that the drug is distributed through total body water. Ethosuximide is not bound to plasma proteins, and it is present in cerebrospinal fluid, saliva, and tears at concentrations similar to those found in plasma (5,25, 26, 27, 28).

Ethosuximide penetrates rapidly the blood-brain barrier. In dogs, the half-life of entry of unchanged drug into the cerebrospinal fluid has been estimated at about 4 to 5 minutes, and this implies that after 20 to 30 minutes, the concentration in the cerebrospinal fluid reaches values identical to those found in plasma (22,29). By comparison, half-lives of entry into the cerebrospinal fluid have been found to be in the order of 3 minutes for diazepam, 12 minutes for valproic acid, 16 minutes for phenobarbital, 17 minutes for phenytoin, 18 minutes for carbamazepine, and 43 minutes for primidone (29).

Studies using unlabeled ethosuximide in rats indicate even distribution throughout the body, except for the adipose tissue, in which concentrations are only about one-third of those reached in plasma, brain, and other tissues (30). Ethosuximide also shows a uniform distribution in discrete brain regions, without significant differences in concentration among the cerebral cortex, the midbrain, the cerebellum, and the pons medulla (31).

Ethosuximide crosses the human placenta and is found in neonatal plasma at concentrations similar to those observed in the mother (45). Ratios of breast milk to plasma concentration are in the order of 0.8 to 0.9 (32, 33, 34, 35, 36). Koup et al. (32) estimated that if a nursing infant receives 200 to 600 mL of milk from a mother who has a plasma ethosuximide concentration of 64 µg/mL, the daily dose of ethosuximide ingested by the infant would be 13 to 38 mg. Rane and Tunell (48) reported that the plasma ethosuximide concentration in a suckling infant during the first 5 months of age was about 30% of the maternal plasma concentration. In a separate study, nursed infants had serum ethosuximide concentrations of 15 to 40 µg/mL, that is, ~50% of the value found in their mothers (34).

Elimination and Excretion

After single oral doses in adults, ethosuximide is eliminated with mean plasma half-lives of 40 to 60 hours (6,20,21,30). Total body clearance, which is about 0.01 l/kg-1/hr (11,20), is considerably lower than liver blood flow, a finding indicating that ethosuximide does not undergo a significant first-pass effect and follows restrictive, flow-independent elimination. Plasma ethosuximide half-lives determined after discontinuation of a multiple-dose regimen are similar to those recorded after a single dose (8), whereas total body clearance may decrease slightly during repeated doses, probably because of a reduction in biotransformation rate (6). Although ethosuximide may induce microsomal enzyme activity in rodents (37,38), there is no evidence that enzyme induction occurs in patients receiving long-term treatment (11). Autoinduction of ethosuximide metabolism has been described in rats (14), but it does not seem to occur in humans (6,7).

The half-life of ethosuximide is generally shorter and clearance is generally higher in children than in adults. Mean half-lives of 29 to 39 hours (range, 15 to 68 hours) have been reported in children aged 5 to 15 years (3,4,39). Preliminary data suggest that half-lives in newborns are within the range reported for children (32,33).

Steady-State Plasma Concentrations and Relationship with Dosage

An interval of 7 to 12 days is required for plasma ethosuximide concentrations to reach steady-state conditions after a dosage adjustment (5,8,40). Because of the slow rate of elimination, daily fluctuations in plasma ethosuximide concentration during repeated doses are relatively minor. Therapeutic levels can be maintained throughout a 24-hour period even with once-daily administration (5,8,41,42), but a twice-daily regimen is more frequently used to minimize gastrointestinal side effects potentially associated with large individual doses (42).

In most patients, daily doses of 15 to 40 mg/kg are required to achieve steady-state plasma ethosuximide concentrations within the commonly quoted optimal range of 280 to 700 µmol/L (40 to 100 µg/mL), but the interindividual variation in plasma levels among patients receiving


the same dose is considerable, and dosage needs to be titrated to meet individual needs (40,43, 44).


FIGURE 67.3. Relationship between steady-state serum ethosuximide concentration and dosage in two patients with absence seizures. The solid line shows the relationship anticipated from the lowest concentration-dose data pair if one assumes linear kinetics. The concentration increases disproportionately with increasing daily dosage. (From Bauer LA, Harris C, Wilensky AJ, et al. Ethosuximide kinetics: possible interaction with valproic acid. Clin Pharmacol Ther 1982;31:741-745, with permission.)

Information on the relationship between plasma ethosuximide concentration and daily dosage is somewhat controversial. In a carefully controlled study in 20 adult volunteers, Goulet et al. (8) found that a dosage increase from 500 to 750 mg/day produced a 50% increment in plasma concentration, a finding suggesting a linear relationship between plasma levels and dose. Smith et al. (44), however, reported that in some patients increments in ethosuximide dosage resulted in a disproportionate increase in plasma drug concentration. In a retrospective survey, Bauer et al. (6) also found similar evidence of nonlinear (saturation) kinetics in seven of 10 patients studied at different dosage levels (Figure 67.3). Overall, these data suggest that in individual patients the metabolism of ethosuximide may become saturated within the therapeutic dosage range. This can result in a disproportionately large rise in plasma ethosuximide concentration when dosage is increased, but this phenomenon appears to be less consistent and less important clinically compared with that observed with phenytoin.

Influence of Developmental Factors, Pregnancy, and Disease States on Plasma Ethosuximide Concentrations

Plasma ethosuximide concentrations tend to be lower in children than in adults receiving comparable weightadjusted doses (40,43,45). In a large epidemiologic study, the ratio between ethosuximide concentrations (µg/mL) and daily dosage (mg/kg/day) was found to be 2.23±0.15 in children aged 2.5 to 10 years (n = 48) compared with 3.14±0.15 in older children (≤15 years of age, n = 79) and 3.34±0.15 in patients aged 16 to 34 years (n = 71) (45). There is also some evidence that plasma ethosuximide concentrations decrease during pregnancy and return to baseline after delivery (32,34,35,46), but this requires confirmation.

Because ethosuximide is eliminated largely by hepatic metabolism, its elimination would be expected to be impaired in patients with liver disease, although this has not been formally investigated. Patients with impaired renal function may also exhibit some decrease in ethosuximide clearance, but the effect of renal disease is not anticipated to be marked.

Ethosuximide is not bound to plasma proteins, and it is removed relatively efficiently by hemodialysis. It has been estimated that the half-life of ethosuximide drops to about 3 to 4 hours during dialysis, and ~50% of the drug in the body can be removed during a 6-hour dialysis interval (47). In an epileptic girl undergoing peritoneal dialysis, increasing the daily duration of dialysis caused a decrease in plasma ethosuximide and phenobarbital concentrations that may have contributed to precipitating seizure activity (48). Patients receiving dialysis and who are treated with ethosuximide may need supplemental doses at the beginning or at the end of the dialytic procedure, and hemodialysis could be useful in cases of drug overdose.

Pharmacokinetic Drug Interactions

Effect of Other Drugs on the Kinetics of Ethosuximide

Because valproic acid is frequently combined with ethosuximide, the possibility of a pharmacokinetic interaction between these drugs has been repeatedly investigated. In rats, valproic acid may increase the brain concentration of concurrently administered ethosuximide, possibly by inhibiting its metabolism (49). In epileptic patients, one report indicated that valproic acid may increase plasma ethosuximide levels (50), but the opposite effect has also been described (45,51). Of two studies that formally evaluated this interaction in normal volunteers, one reported no change in ethosuximide kinetics after addition of valproic acid at a dose of 500 mg daily (6), whereas the other documented a small but statistically significant increase in plasma ethosuximide concentration after administration of a larger dose of valproate (800 to 1,600 mg/day) (21). In practice, the combination of valproic acid and ethosuximide is often clinically beneficial. Rowan et al. (52) found that absence seizures refractory to either drug may respond remarkably well to the combination: the mechanism underlying


this favorable response does not appear to involve major pharmacokinetic changes, and it is probably pharmacodynamic (49).

Because ethosuximide is eliminated primarily by oxidative biotransformation, it is not surprising that its rate of metabolism is accelerated by enzyme-inducing anticonvulsants, such as phenobarbital, primidone, phenytoin, and carbamazepine (11,16, 17, 18). In one study, Giaccone et al. (18) found that ethosuximide clearance in 10 epileptic patients receiving long-term therapy with phenobarbital, phenytoin, or carbamazepine was about 65% higher than that observed in 12 study subjects who did not receive comedication. In an earlier study in six healthy volunteers, a carbamazepine dose as low as 200 mg daily was sufficient to decrease plasma ethosuximide concentrations by 17% and to shorten its half-life from 54 to 45 hours (17).

Information about the effect of nonanticonvulsants on ethosuximide kinetics is scant. Coadministration of rifampicin increases ethosuximide clearance in normal volunteers, presumably because of enzyme induction (16). Another drug used for the treatment of tuberculosis, isoniazid, may cause ethosuximide intoxication by inhibiting its metabolism (53), but evidence for this interaction is inconclusive. In rats, ethosuximide clearance can be markedly reduced by certain enzyme inhibitors, including relatively selective inhibitors of cytochrome CYP3A such as triacetyloleandomycin (15,54). If ethosuximide is metabolized by the same cytochrome in humans, similar interactions would be expected to occur in a clinical setting.

Effect of Ethosuximide on the Kinetics of Other Drugs

Contrary to findings obtained in rodents (14,37, 38), no evidence indicates that therapeutic doses of ethosuximide cause enzyme induction in humans (11,55). Ethosuximide is not bound to plasma proteins, and therefore competition with other drugs at plasma protein binding sites cannot occur. Because of these properties, ethosuximide has a lower interaction potential than many other anticonvulsants.

In rats, ethosuximide was found to increase the brain concentration of concurrently administered valproic acid, but this effect was observed only at neurotoxic doses (49). In epileptic patients, sporadic reports suggest that ethosuximide may increase the concentration of phenytoin (56,57) and phenobarbital derived from primidone (58), but the clinical significance of these potential interactions is unclear.


  1. Zimmerman FT, Burgmeister BB. A new drug for petit mal epilepsy. Neurology1958;8:769-775.
  2. Edwardson JM, Dean PM. Inhibition of alloxan-induced hypercalcemia by compounds of similar molecular structure. Biochem Pharmacol1992:44:2111-2115.
  3. Chang T. Ethosuximide: biotransformation. In: Levy R, Mattson R, Meldrum B, Penry JR, et al., eds. Antiepileptic drugs,3rd ed. New York: Raven Press, 1989:679-683.
  4. Buchanan RA, Fernandez L, Kinkel AW. Absorption of elimination of ethosuximide in children. J Clin Pharmacol1969;7: 213-218.
  5. Buchanan RA, Kinkel AW, Smith TC. The absorption and excretion of ethosuximide. Int J Clin Pharmacol1973;7:213-218.
  6. Bauer LA, Harris C, Wilensky AJ, et al. Ethosuximide kinetics: possible interaction with valproic acid. Clin Pharmacol Ther1982;31:741-745.
  7. Glazko AJ. Antiepileptic drugs: biotransformation. metabolism, and serum half-life. Epilepsia1975;16:367-391.
  8. Goulet GR, Kinkel AW, Smith TC. Metabolism of ethosuximide. Clin Pharmacol Ther1976;20:213-218.
  9. Millership JS, Mifsud J, Collier PS. The metabolism of ethosuximide. Eur J Drug Metab Pharmacokinet1993;18:349-353.
  10. Pisani F, Bialer M. Chemistry and biotransformation. In: Levy RH, Mattson RH, Meldrum BS, eds. Antiepileptic drugs,4th ed. New York: Raven Press, 1995:655-658.
  11. Eadie MJ, Vajda FJE. Older anticonvulsants continuing in use but with limited advances in knowledge. In: Eadie MJ, Vajda FJE, eds. Handbook of experimental pharmacology,vol 13: Antiepileptic drugs, pharmacology and therapeutics,Berlin: Springer-Verlag, 1999:189-228.
  12. Horning MO, Stratton J, Nowlin DJ, et al. Metabolism of 2-ethyl-2-methyl succinimide (ethosuximide) in the rat and human. Drug Metab Dispos1973;3:569-576.
  13. Pettersen JE. Urine metabolites of 2-ethyl-2 methyl-succinimide (ethosuximide) studied by combined gas chromatography mass spectrometry. Biomed Mass Spect1978:5:601-603.
  14. Bachmann KA, Jahn D, Yang C, et al. Ethosuximide disposition kinetics in rats. Xenobiotica1988;18:373-380.
  15. Bachmann KA, Chu CA, Grear V. In vivo evidence that ethosuximide is a substrate for cytochrome P450IIIA. Pharmacology1992;45:121-128.
  16. Bachmann KA, Jauregui L. Use of single sample clearance estimates of cytochrome P450 substrates to characterize human hepatic CYP status in vivo. Xenobiotica1993;3:307-315.
  17. Warren JW, Benmaman JD, Wannamaker BB, et al. Kinetics of a carbamazepine-ethosuximide interaction. Clin Pharmacol Ther1980;28:646-651.
  18. Giaccone M, Bartoli A, Gatti G, et al. Effects of enzyme inducing anticonvulsants on ethosuximide pharmacokinetics in epileptic patients. Br J Clin Pharmacol1996;41:575-579.
  19. Villen T, Bertilsson L, Sjoqvist F. Nonstereoselective disposition of ethosuximide in humans. Ther Drug Monit1990;12:514-316.
  20. Eadie MJ, Tyrer JH, Smith JA, et al. Pharmacokinetics of drugs used for petit mal absence epilepsy. Clin Exp Neurol1977;14: 172-183.
  21. Pisani P, Narbone MC, Trunfio C, et al. Valproic acid-ethosuximide interaction: a pharmacokinetic study. Epilepsia1984;25: 229-233.
  22. EI Sayed MA, Loscher W, Frey HH. Pharmacokinetics of ethosuximide in the dog. Arch Int Pharmacodyn Ther1976;234:180-192.
  23. Patel IH, Levy RH. Pharmacokinetic properties of ethosuximide in monkeys. II. Chronic intravenous and oral administration. Epilepsia1975;16:717-730.
  24. Patel IH, Levy RH. Bauer TG. Pharmacokinetic properties of ethosuximide in monkeys. I. Single dose intravenous and oral administration. Epilepsia1975;16:705-716.
  25. Horning MG, Brown L, Nowlin J, et al. Use of saliva in therapeutic drug monitoring. Clin Chem1977;23:157-164.
  26. McAuliffe JJ, Sherwin AL, Leppik IE, et al. Salivary levels of anticonvulsant: a practical approach to drug monitoring. Neurology1977;27:409-413.



  1. Piredda S, Monaco F. Ethosuximide in tears, saliva and cerebral fluid. Ther Drug Monit1981;3:321-323.
  2. Loscher W. A comparative study of the protein binding of anticonvulsant drugs in serum of dog and man. J Pharmacol Exp Ther1979;208:429-435.
  3. Loscher W, Frey HH. Kinetics of penetration of common antiepileptic drugs into cerebrospinal fluid. Epilepsia1984;25: 346-352.
  4. Dill WA, Peterson L, Chang T, et al. Physiologic disposition of alpha-methyl-alpha- ethylsuccinimide (ethosuximide; Zarontin) in animals and in man. In: Abstracts of papers, 149th national meeting, American Chemical Society, Detroit, Michigan.Washington, DC: American Chemical Society, 1965:30N.
  5. Patel IH, Levy RH, Rapport RL. Distribution characteristics of ethosuximide in discrete areas of rat brain. Epilepsia1977;18: 533-541.
  6. Koup JR, Rose JQ, Cohen ME. Ethosuximide pharmacokinetics in a pregnant patient and her newborn. Epilepsia1978;19: 535-539.
  7. Kaneko S, Sato T, Suzuki K. The levels of anticonvulsant in breast milk. Br J Clin Pharmacol1979;7:624-627.
  8. Kuhnz W, Koch S, Jacob S, et al. Epileptic women during pregnancy and lactation: placental transfer, serum concentration in nursed infants and clinical status. Br J Clin Pharmacol1984;18: 671-677.
  9. Rane A, Tunell R. Ethosuximide in human milk and in plasma of a mother and her nursed infant. Br J Clin Pharmacol1981;12: 855-858.
  10. Hagg S, Spigset O. Anticonvulsant use during lactation. Drug Saf2000;22:425-440.
  11. Orton TC. Nicholls PJ, Effect in rats of subacute administration of ethosuximide, methsuximide ad phensuximide on hepatic microsomal enzymes of porphyrin turnover.Biochem Pharmacol1972;21:2253-2261.
  12. Stevenson IH, O'Malley K, Shepherd AMM. Relative induction potency of anticonvulsant drugs. In: Richens A, Woodford FP, eds. Anticonvulsant drugs and enzyme induction.Oxford: Elsevier, 1976:37-46.
  13. Buchanan RA, Kinkel AW, Turner JL, et al.. Ethosuximide dosage regimens. Clin Pharmacol Ther1976;19:143-147.
  14. Browne TR, Dreifuss FE, Dyken PR, et al. Ethosuximide in the treatment of absence (petit mal) seizures. Neurology1975; 25:515-524.
  15. Colburn WA, Gibaldi M. Use of MULTIDOS for pharmacokinetic analysis of ethosuximide data during repeated administration of single or divided daily doses. J Pharm Sci1978;67:574-573.
  16. Dooley JM, Camfield PR, Camfield CS, et al. Once daily ethosuximide in the treatment of absence epilepsy. Pediatr Neurol1990;6:38-39.
  17. Sherwin AL. Ethosuximide: clinical use. In: Levy RH, Dreifuss FE, Mattson RH, et al., eds. Antiepileptic drugs,3rd ed. New York: Raven Press, 1989:679-698.
  18. Smith GA, McKauge L, Dubety D, et al. Factors influencing plasma concentrations of ethosuximide. Clin Pharmacokinet1979;4:38-52.
  19. Battino D, Cusi C, Franceschetti S, et al. Ethosuximide plasma concentrations: influence of age and associated concomitant therapy. Clin Pharmacokinet1982;7:176-180.
  20. Eadie MJ, Lander CM, Tyrer JH. Plasma drug level monitoring in pregnancy. Clin Pharmacokinet1977;2:427-436.
  21. Marbury TC, Lee CC, Perchalski RJ. Hemodialysis clearance of ethosuximide in patients with chronic renal failure. Am J Hosp Pharmacol1981;38:1757-1760.
  22. Marquardt ED, Ishisaka DY, Batra KK, et al. Removal of ethosuximide and phenobarbital by peritoneal dialysis in a child. Clin Pharmacol1992;11:1030-1031.
  23. Bourgeois BFD. Combination of valproate and ethosuximide: antiepileptic and neurotoxic interaction. J Pharmacol Exp Ther1988;247:1128-1132.
  24. Mattson RH, Cramer JA. Valproic acid and ethosuximide interaction. Ann Neurol1980: 583-584.
  25. Flachs H, Wurtz-Iorgensen A, Gram L, et al. Sodium di-n-propylacetate: its interaction with other antiepileptic drugs. In: Schneider H, Janz D, Gardner-Thorpe C, et al., eds.Clinical pharmacology of antiepileptic drugs.Berlin: Springer-Verlag, 1975: 163-172.
  26. Rowan AJ, Meijer JWA, de Beer-Pawlikowski N, et al. Valproate ethosuximide combination therapy for refractory absence seizures. Arch Neurol1983;40:797-802.
  27. Van Wieringen A, Vrijlandt CM. Ethosuximide intoxication caused by interaction with isoniazid. Neurology1983;33: 1227-1228.
  28. Bachmann KA, Madhira MS, Rankin GO. The effect of cobalt chloride, SKF-525A, and N- (3,5-dichlorophenyl) succinimide on in vivohepatic mixed function oxidase activity as determined by single-sample plasma clearances. Xenobiotica 1992;22:27-31.
  29. Gilbert JC, Scott AK. Galloway DB, et al. Ethosuximide: liver enzyme induction of D-glucaric acid excretion. Br J Clin Pharmacol1974;1:249-252.
  30. Dawson GW, Brown HW, Clark BG. Serum phenytoin after ethosuximide. Ann Neurol1978;4:583-584.
  31. Lander CM, Eadie MJ, Tyrer J. Interactions between anticonvulsants. Proc Aust Assoc Neurol1975;12:111-116.
  32. Schmidt D. J Neurol1975;209:115-123.