Clinical Pharmacokinetics 1: 161-188 (1976) © ADIS Press 1976

Clinical Pharmacokinetics of Anticonvulsants

E. F. Hvidberg and M. Dam Clinical Pharmacology Research Unit, Rigshospitalet, University of Copenhagen, and Department of Neurology, Hvidovre Hospital (University Hospital), Copenhagen

Summary

Anticonvulsant therapy was among the first areas to benefit from clinical pharmacokinetic studies. The most important advantage is that the frequent interindividual variation in the plasma level/dose ratio for these drugs can be circumvented by plasma level monitoring. For several anticonvulsants the brain concentration is shown to parallel the plasma concentration. Phenytoin (diphenylhydantoin) is still the most important anticonvulsant and the one for which kinetics have been thoroughly investigated in man. These investigations have revealed several reasons for the wellknown difficulties in using this drug clinically. The absorption rate and fraction are very much dependent on the pharmaceutical preparation, and changes of brand may alter the plasma level of phenytoin in spite of unaltered dose. The elimination capacity is saturable causing dose dependent kinetics, which again means disproportional changes in plasma level with changes in dose. Great individual variations exist in the rate of metabolism, and several pharmacokinetic drug interactions are known. As an optimum therapeutic plasma concentration range has been established monitoring plasma levels must be strongly advocated. Interpretation of plasma levels in uraemic patients must take into account decreased protein binding of the drug. Carbamazepine is probably as effective as phenytoin. The elimination is a first order process, but the rate of metabolism increases after a few weeks' treatment. An active metabolite (epoxide) may be the cause of some side-effects. Combined treatment with other anticonvulsant drugs decreases the halflife and more frequent dosing may be necessary. An optimum therapeutic concentration range has been suggested and plasma monitoring is advocated, along with that of the active metabolite, the epoxide. Phenobarbitone is still much used but its kinetics have been investigated to a lesser extent. The main problem is the variability in the rate of elimination. In children the half-life of phenobarbitone is only half of that in adults. An optimum therapeutic plasma range has been established and monitoring is recommended. Primidone may have an anticonvulsant activity in itself. but its main metabolite is phenobarbitone. The relatively rapid elimination of primidone is offset by the long halflife of phenobarbitone. An optimum therapeutic range has been suggested, but plasma level monitoring must include determination of phenobarbitone. Ethosuximide. The clinical pharmacokinetics of this important petit mal anticonvulsant is not well known. It has a relatively long halflife (in adults 2 to 3 c/Qys; in children shorter). An optimum therapeutic range has been suggested, and routine monitoring of plasma levels may be recommended. Diazepam exerts a rapid anticonvulsant activity when the plasma concentration exceeds approximately 500ng/ml after intravenous injection. The kinetic pattern is complex in man.

162

Clinica l Pharmacokinetics of Anticonvulsants

Oonazepam. The clinical pharmacokinetics are still not fully investigated but a th era· peutic range has been suggested. Monitoring of plasma levels may be carried out 0 11 special o ccasions. Di-n-propylacetic acid (valproic acid) is a n ew anticon vulsant, which is kinetically not sufficiently in vestigated in man. With a half-life of about 10 hours greater fluctuatiollS il1 the plasma concentration may be seen. Plasma monitoring cannot yet be recommended as a routine procedure.

Treatment with anticonvulsant drugs was one of the first clinical areas to benefit from human pharmacokinetic investigations, starting with the fundamental work in the fifties and early sixties by Buchthal and his group. Since then, vast volumes of literature have been published on the kinetics of these drugs. The reason for the success of plasma level monitoring in for example epileptic patients, is that individual variations in the plasma level/dose ratio can be circumvented by measuring plasma concen trations (e .g. see Eadie, 1976). Therefore, the relation between plasma levels and effect should in particular be able to be evaluated, but controlled and reliable studies are, in contrast to pure kinetic studies , unfortunately quite scarce. Among many other problems in the clinical pharmacokinetic properties of anticonvulsants are the following:

1) Is the plasma concentration an acceptable measure of the concentration in the brain? Several recent papers have question positively.

answered

this

2) Do the epileptic disorders by themselves influence the pharmacokinetics of the drugs? So far no evidence to support this notion has been produced. 3) In the treatment of epilepsy , combined therapy is fairly commonly used . How are the

kinetics of the single drugs influenced by the other drugs given simultaneously and to what extent is this of clinical significance? Much more work has to be done in this field.

In the present article we do not intend to review all kinetic studies on anticonvulsant drugs

carried out in humans, nor to review any kinetic studies in animals. Rather, we have tried to indicate areas and situations in which the existing kinetic information may be applied after a critical evaluation. A more mathematical approach has not been intended. We have confined ourselves to the most commonly used antiepileptic drugs . However, a few of the newer ones will also be discussed briefly. To the extent that reliable information is available, each drug has been discussed along a similar general scheme.

1. Phenytoin (Diphenylhy dantoin) Phenytoin is still one of the most commonly used anticonvulsant drugs . The amount of literature concerned with the clinical pharmacokinetics of phenytoin is so extensive that a complete evaluation will not be possible within the scope of this review. Fortunately, a series of very comprehensive reviews of the literature up till abo ut 1972 is available (see Woodbury et aI., 1972). The following provides brief statements of the information contained in this review followed by a discussion of more recent publications. The data are summarised in table I.

1.1 Basic Human Pharmacokinetics

1.1.1 Absorption Phenytoin is sparingly soluble in water and has a pKa value around 9. It is mainly absorbed from the proximal part of the small intestine. The rate

Clinical Pharmacokinetics of Anticonvulsants

Table I. Clinical pharmacokinetic profile of phenytoin Property

Notes

Absorption rate

Slow, variable (Mostly dependent on the pharo maceutical preparation)

Absorption fraction

20 to 90% (Mostly dependent on the pharo maceutical preparation)

Apparent volume of distribution 0 .5 to 0.8L/kg; 5 to 6L/kg for the unbound phenytoin Distribution to CNS, CSF

Fast: CSF conc. parallels free plasma conc.

Protein·binding

87 - 93%. Lower in uraemia

Elimination : biotransformat ion

'\195%

Elimination : renal (unchanged)

1 to 5%

Apparent plasma half· life

Dose-dependent (8 to 60h) . Great inter· and intra-individual variation

Plasma clearance

"vO.02L/kg/h at linear kinetics (van der Kleijn, 1975)

Pattern of elimination kinetics

Non-linear

Kinet ic model applicable

Two compartment at linear ki netics

Metabol ites

HPPH ( inact ive)

Steady-stat e/dose ratio

Many fold variation

T herapeutic plasma range

10 to 20/lg/ml

Dose schedule

Twice daily (adults) . Once-a-day schedule has been proposed

Special featu res

Many drug interactions

~

linear

of absorption is slow and variable; peak concentrations of phenytoin are attained from 4 to 24 hours after oral administration. The systemic

163

bioavailability of phenytoin from tablets from different makers differs greatly (20 to 90%). Difference in particle size is the crucial determinant (Lund, 1974a), but other factors may also be of influence (Johannessen and Strandjord, 1975a). Thus, the sodium salt seems to be more readily absorbed. As a result, change of brand or even batch, might cause considerable changes in the plasma level of phenytoin (Alvan et aI. , 1975; Pentikiiinen et aI. , 1975 ; Stewart et aI. , 1975).

1.1.2 Distribution The apparent volume of distribution is about 0.5 to 0.8L/kg. Phenytoin is bound to plasma protein about 93% (Lunde et al., 1970)_ The volume of distribution for the unbound fraction is about 5L/kg (Odar-CederiOf and Borg~, 1974)_ Many conditions may influence this binding (see sections 1.2, 1.3). Although it is a matter of dispute how great the clinical significance is, interest in measuring the free (unbound) phenytoin fraction is increasing. At least five approaches have been made: 1) Direct measurements, but in many places the methods (e.g. ultrafiltration, dialysis etc.) are apparently not yet suitable for routine purposes. 2) A kind of direct estimation was suggested by Christiansen et al. (1975) utilising the phenomeilon that venous stasis increase the plasma protein concentration of a blood sample. This idea remains to be further examined. 3) Measuring the phenytoin concentration in CSF gives a good estimation of the free concentration (see review by Johannessen and Strandjord, 1975a), but can, of course , not be applied to daily clinical practice. 4} Analysing phenytoin in saliva seems to provide an easy expression of the free concentration (Bochner et aI., 1974; Troupin and Friel, 1975; Cook et aI., 1975). Also, carbamazepine , phenobarbitone and primidone can be measured by this method, and protein binding problems (low albumin, uraemia, drug interactions etc.) can be overcome.

Clinical Pharmacokinet ics of Anticonvulsants

5) Binding of phenytoin (and phenobarbitone) to erythrocytes is rather constant and unaffected by e.g. uraemia (see later). It has therefore been suggested that red blood cell/plasma concentration ratio may serve as a screening method for abnormal plasma binding (Glazko, 1973; Borondy et al., 1973; Kurate and Wilkinson, 1974; Ehrnebo and Odar-CederlOf, 1975). The value of all the mentioned methods for measuring the free fraction remains to be proven in clinical practice. Furthermore, general use of methods for the determination of the free fraction of phenytoin in antiepileptic treatment may probably not be relevant. Although some studies fInd a considerable interindividual variation in the degree of protein binding of phenytoin (Hooper et al., 1974a; Bochner et al., 1974), others have observed only a minor variation (Lunde et al., 1970; Lund et al., 1972). In a recent investigation, only a 2-fold interindividual variation was demonstrated and the individual degree of protein binding for phenytoin was reproducible at different times (Barth et al., 1976). This means that monitoring total plasma concentration should be sufftcient, provided the patients do not suffer from renal or hepatic diseases (see section 1.3). The distribution of phenytoin to the central nervous system and to CSF has recently been studied in humans. The concentration in CSF corresponds fairly well to the non-protein bound fraction (see above), although some variations are noted. By examining surgical biopsies of the brain, both Vajda et al. (1974) and Houghton et al. (1975b) found excellent correlations between the concentrations in plasma and brain. This is a strong indication, although not proof, of a correlation with the concentration at the receptor-site. 1.1.3 Metabolism In man, phenytoin is metabolised predominantly to its parahydroxy metabolite (HPPH) , which in turn is conjugated to HPPH-glucuronide. The latter metabolite is present in plasma with a

164

concentration about 10 times higher than HPPH, while the ratio of HPPH to DPH is as low as 0.06 (Albert et aI. , 1974). The p-hydroxylation is the major metabolic pathway and rate-limiting step (see below). 75% of an intravenous dose of 250mg can be recovered in the urine as HPPH and HPPH-glucuronide, but during continuous treatment the recovery may be lower (Glazko and Chang, 1972). Several other metabolites have been identified (Glazko, 1973), but so far none of them can be considered active. One possible exception may be an intermediate formation of epoxides, which might be toxic to the hepatocytes, but no relevant clinical information on this problem is available as yet. 1.1.4 Elimination Kinetic Pattern Although known since the late sixties, it is only during the last yew years that it has been more widely accepted that phenytoin elimination is subject to dose dependent kinetics which is within the therapeutic range in most patients. The realisation of this phenomenon explains the difftculties in adjusting treatment with phenytoin to the right level. The reason for non-linear kinetics is the partial saturation of the p-hydroxylation pathway (Remmer et al. , 1969; Arnold and Gerber , 1970; Atkinson and Shaw, 1973; Houghton and Richens, 1974; Richens and Dunlop, 1975a). Because of its saturation kinetics it is not possible to precisely state the half-life of phenytoin in plasma, as this will vary with the dose. However, for practical purposes an average half-life of about 20 to 30 hours can be given, but with a many-fold inter- and intra-individual variation. The present review does not intend to cover the many biochemical and theoretic-kinetic aspects of the saturation phenomenon. The reader is referred to several comprehensive and recent reviews (e.g. Woodbury et al., 1972; van der Kleijn et aI., 1975). Values for Km and Vmax can be estimated and the steady-state level predicted on the basis of equations summarised by for example, van der Kleijn et al. (1975).

Cl ini cal Pharmacokinetics of Anticonvulsants

1.2 Influence Kinetics

of Physiological

States on

1.2.1 Sex Sex does not appear to influence plasma concentration (Hooper et al., 1974c). 1.2.2 Age A marked increase of phenytoin clearance in older patients (>65 years) , after both oral and intravenous administration correlates inversely with phenytoin binding and plasma albumin, both of which are reduced in the elderly (Hayes et ai., 1975). Increased clearance, which results in a lower total concentration, and increased free fraction of phenytoin seem to counterbalance each other and the dose should therefore not be altered in the elderly. When phenytoin is given in proportion to body weight in children the plasma level is found to be lower than in adults, probably because of a higher rate of excretion (Svensmark and Buchthal, 1964). When related to age, children below 11 to 12 years tend to require higher doses than older children or adults (Berlet, 1975). Hooper et al. (1974b) suggested that this variation was due to competition for hydroxylation enzymes between phenytoin and steroids; the output of the latter increasing markedly with puberty. The phenomenon may, however, be related to the greater ratio of hepatic size to body weight in small children, this ratio progressively falling with body growth. Phenytoin clearance is correlated to liver size in adults (Bach et ai., to be published). 1.2.3 Pregnancy During pregnancy, plasma clearance of phenytoin is often more than doubled, resulting in plasma concentrations below the expected range. In some patients the change is observed by the second month of pregnancy (Dam et al., 1976). As protein binding does not differ significantly between pregnant and non-pregnant females (Hooper et al., 1974a) increased metabolism of phenytoin induced by the increased hormone

165

production is suggested (Dam et al., 1976). Frequent plasma monitoring of phenytoin should therefore act as a safe guideline for phenytoin treatment during pregnancy .

1.2.4 Neonates The concentration of phenytoin in plasma samples obtained from the umbilical artery or vein is almost identical with the respective maternal plasma (Mirkin, 1971). The plasma disappearance rate of phenytoin in newborns of mothers treated long-term with phenytoin is in the same range as seen in adults (Hoppel et ai., 1975). The disappearance rate correlates well with the metabolite formation rate (Rane, 1974), suggesting an induction in utero. Therefore, such data should not be extrapolated to newborn infants of nontreated mothers (Hoppel et al., 1975). The plasma protein binding in neonates is lower than in adults (Ehrnebo et al., 1971). 1. 2. 5Breast Milk Breast milk and colostrum from women on chronic anticonvulsant therapy contain concentrations of phenytoin which are significantly below those of maternal plasma. A limited transport capacity of the mammary glands for phenytoin (Mirkin, 1971) is less likely, as acidic substances (e.g. phenytoin) have a milk/blood ratio lower than one (Rasmussen, 1971).

1.2.6 Genetic and Ethnic Factors The great interindividual variation in the rate of metabolism of phenytoin, has in twin studies, been shown to be determined predOminantly by genetic factors (Andreasen et al., 1973). Accumulation of unmetabolised phenytoin caused by insufficient p-hydroxylation has been identified in two generations of a family, suggesting a genetic enzymatic defiCiency to be the cause (Kutt et ai., 1964c). An ethnic difference was suggested as the cause of lower phenytoin plasma concentrations in 45 in-patients from Greenland compared with patients from Denmark (Dam et al., 1974).

Clinical Pharmacokinetics of Anticonvulsants

1.2.7 Other Factors Body weight, height and probably other factors may influence the plasma levels of phenytoin, but as pointed out by Houghton et aI. (1975a), genetic factors and the effect of the saturation kinetics are much more important in determining steady state concentrations. 1.3 Influence of Disease States on Kinetics

1.3.1 Renal Disease The protein binding of phenytoin is decreased in uraemic patients (Odar-CederlOf et aI., 1970; Reidenberg et aI., 1971; Andreasen, 1973; OdarCederiOf and Borg~, 1974; Hooper et aI., 1974a). This could be due to qualitative differences in plasma albumin in these individuals (Reidenberg et aI., 1971; Shoeman et aI., 1973), but recent investigations indicate the presence of a binding inhibitor which is irreversibly bound to albumin at physiological pH and which affects at least some drugs in serum from uraemic patients (Sjoholm et aI., 1976). Phenytoin binding is dependent on albumin concentration (the Boston Collaborative Drug Surveillance Program, 1973) and on the degree of renal failure (e.g. measured by the creatinine clearance), but not on total protein concentration (Olsen et aI., 1975). Patients with chronic uraemia metabolise phenytoin more rapidly, as the plasma concentration is low, half-life decreases and HPPH concentration is markedly elevated (Mellk et aI., 1970; Odar-CederiOf and Borg~, 1974). Although the free fraction is increased, it is not necessary to reduce the dosage, since the absolute concentration of unbound drug is unchanged due to a compensatory increase in the apparent volume of distribution and plasma clearance (Odar-CederlOf and Borgg, 1974; Gugler et aI., 1975). 1.3.2 Liver Disease In patients with liver disease, an accumulation of unmetabolised drug and low output of the metabolites are demonstrable with commonly used

166

daily dosages (Kutt et aI., I 964b). Two patients suffering from severe acute viral hepatitis showed small decreases in plasma clearance, while three other patients in the recovery phase of their illness showed 20 to 40% increase in phenytoin clearance over values from the control phase (Blaschke et aI., 1975). A major problem in attempts to relate elimination rate to liver function is that the latter is seldom quantitated. Protein binding may also change in liver disease. Hooper et aI. (1974a) found binding to be decreased, correlating with changes in albumin and bilirubin levels in plasma, but only 3 of 8 patients with alcoholic liver disease had decreased plasma protein binding of phenytoin. Thus binding of phenytoin in plasma of most patients with alcoholic liver disease is not substantially impaired (Affrime and Reidenberg, 1975). The net effect of impaired liver function on phenytoin kinetics may reflect a balance between an increased fraction of unbound phenytoin and a possible reduction in the capacity of the liver microsomal enzyme system to metabolise drugs. The complex pattern of the influence of liver disease on the disposition of phenytoin is, however, not fully clarified and no safe guideline can yet be recommended, except for monitoring of plasma levels, together with close clinical observations of the patient.

1.3.3pH In severe intoxication, plasma pH may affect the distribution of phenytoin as the plasma protein binding increases as the pH increases (pruitt et aI., 1975). This observation should be of Significance for the treatment of phenytoin intoxication. 1.4 Implications of Clinical Pharmacokinetic Properties to Therapy

1.4.1 Relation of Activity to Plasma Level Therapeutic Effect: A therapeutic, or better an optimum plasma range (10 to 20 J..Lg/ml, or 40 to

Clinica l Pharmacokinetics of Anticonvulsants

80J..lmol/ L total plasma concentration) is now well established, but for the individual patient the optimum plasma concentration may vary and even be found outside this range (Lund , 197 4b). Buchthal and Svensmark (1960) in a prospective study found that convulsions in most patients with severe grand mal epilepsy were controlled or reduced in number with phenytoin blood levels about 15 J..Ig/ ml. Similarly, Lund (1973) in a retrospective study found the mean plasma levels in patients without seizures to be Significantly higher (12.6 ± 8.8 J..Ig/ml) than the levels in patients with seizures (8.6 ± 4.6 J..Ig/ml). In a prospective study, the mean phenytoin plasma level was 6.1 ± 2 .9 J..Ig/ml during the first year with a mean number of seizures of 5.8 per patient. The corresponding figures for the second year were 11 .7 ± 3.3 J..Ig/ml and 4.1 seizures , and for the third year 15 .0 ± 2.6 J..Ig/ml and 1.6 seizures per patient per year (Lund , 1974b). Somewhat in contrast , Wilder et al. (1972) reported that clinical response and CSF levels of phenytoin do not correlate very well. One probable reason could be the difficulties of a precise estimation of the low concentrations in CSF. Side-effects: Toxic side-effects are usually not seen with phenytoin plasma levels below 14J..1g/ ml, whereas half of the patients with phenytoin levels of 30J..lg/ml show side-effects (Buchthal and Svensmark, 1960). Nystagmus can appear at a level about 20J..lg/ml , ataxia at 30J..lg/ml and mental changes at 40J..lg/ ml and above (Kutt et al., 1964a).

1.4.2 Dosing Problems Intravenous administration:

Intravenous administration of phenytoin may be used when therapeutic levels must be achieved rapidly (e .g. treatment of status epilepticus or acute cardiac arrhythmias). Slow intravenous injection (10mg/kg body weight of phenytoin during 5 to 10 min) will at once lead to a level of phenytoin between 10 to 20J..lg/ml in nearly all patients. It should be borne in mind that phenytoin cannot be dissolved in most parenterally used solutions and will precipitate if it is diluted. If dilution is necessary, it is

167

recommended that the pharmacist should prepare the solution. However, special solutions for phenytoin are available in some countries (Oelkers et al., 1975). Intramuscular administration: A model for intramuscular administration has recently been proposed (Kostenbauder et al., 1975). However , the absorption of phenytoin is slower after intramuscular administration than after oral ingestion, and the plasma level obtained with the same dose is lower (Olesen and Dam, 1965; Dam and Olesen, 1966; Wilder et al., 1974). If administration of phenytoin is changed from oral to intramuscular, a 50% increase in dosage may be needed to prevent a fall in serum concentration (Wilder and Ramsay , 1976). Comparison of phenytoin plasma concentration in an intravenous (250mg) and intramuscular (500mg) cross-over study indicates that phenytoin administered intramuscularly is absorbed over a period of approximately 5 days (Kostenbauder et al ., 1975). Because of this slow and often erratic absorption, and also because of painful local reactions, intramuscular administration of phenytoin is not recommended in general . Single daily oral dose: After ingestion of a single daily dose of phenytoin during continuous treatment, the fluctuations in phenytoin plasma level during 24 hours are in many cases within a clinically acceptable range without any reduction of the anticonvulsant effect (Strandjord and Johannessen, 1974; Cocks et aI. , 1975). The s lo w rate of absorption, particularly of older preparations (?), may partly be the explanation (Buchanan et aI. , 1972). More episodes of nystagmus have , however , been observed (Buchanan et aI. , 1972). Other problems are involved in this 'single dose problem', and individual differences (also in patient habits) are probably not uncommon . Although patient compliance may be increased with a single daily dose , the decision of whether the daily dose be taken in 1, 2 (or 3) doses must be made separately for each patient. The dose problems for the individual patient can best be solved on the basis of plasma pheny-

168

Clinical Pharmacokinetics of Anticonvulsants

toin monitoring, so this is highly recommended whenever it is possible. It has also been proposed (Richens and Dunlop, 1975a,b) to use a nomogram to predict plasma concentrations in patients recelvmg various doses of phenytoin. This approach has been criticised (Philips et al., 1975; Lund and Alvan, 1975) as being too inexact and its general use instead of plasma monitoring cannot be advocated. Drug interactions: Numerous kinetic interactions have been observed experimentally, but the clinical significance is often a matter of dispute (Kutt, 1975). Only interactions of well established clinical relevance will be mentioned. Dicoumarol (bishydroxycoumarin) has been shown to depress the metabolism of phenytoin (Hansen et al., 1966), which itself induces the metabolism of dicoumarol (Hansen et aI., 1971). Isoniazid, and to a lesser extent p-aminosalicylic acid, inhibit the parahydroxylation of phenytoin in those who are slow isoniazid metabolisers (Kutt et al., 1970; Brennan et al., 1970). Chloramphenicol (Christensen and Skovsted, 1969), disulfrram (Olesen, 1967; Svendsen et al. 1976), sulthiame (Hansen et al., 1968), phenyrarnidol (Solomon and Schrogie, 1967), phenylbutazone (Hansen et aI., 1966), and sulphamethizole (Lurnholtz et al., 1975) also inhibit the metabolism of phenytoin causing higher plasma levels of the latter. Barbiturates and carbamazepine cause lower steady state levels of phenytoin (see section 3.3.3). With any anticonvulsant, an interaction affecting plasma concentrations is not necessarily always harmful. Thus, elevation of a low plasma level of the primary drug may well improve seizure control. On the other hand, a modest decrease in the plasma level of the primary drug may be of no therapeutic consequence (Kutt, 1975).

1.5 Conclusions Phenytoin is still the most important anticonvulsant drug and the one whose kinetics is best

studied. The major kinetic problems are the variability of absorption (biopharmaceutical causes), the dose-dependent kinetics causing disproportional changes in plasma level with small alterations in dosage, and frnally the considerable individual variation in metabolism. All three difficulties are strong indications for monitoring plasma levels in patients, as the ratio between plasma level and effects is in general well established.

2. Other Hydantoins The human, not to say the clinical, pharmacokinetics of hydantoin anticonvulsants other than phenytoin are virtually unknown. Furthermore, no controlled studies of their efficacy seem to have appeared. Mephenytoin, a/butoin and phenacemide: Only for albutoin has a method for plasma assay been developed (Cereghino et aI., 1972). The plasma levels found were about or mostly below 4 Jig/mI, but more elaborate kinetic studies have not as yet been conducted. Ethotoin: Methods for the determination of ethotoin in plasma have recently been developed (Larsen and Naestoft, 1974; Yonekawa et aI., 1975; Miyamoto et al., 1975). However, only few kinetic studies have yet been published. The most important observation is that non-linear kinetics exists within the therapeutic range (Sjo et al., 1975a). This is probably due to a saturation of ethotoin's hepatic metabolism (Naestoft et aI., 1976). The saturation kinetics of ethotoin means that all the kinetic problems involved in the use of phenytoin most probably also concern ethotoin. The half-life of ethotoin (when metabolised by linear kinetics below 8 Jig/mI) is about 5 to 9 hours (Sjo et al., 1975a; Yonekawa et al., 1975). A therapeutic plasma level has not been established, but the usual level seems to lie between 5 and 20Jig/rnl (pre-dose concentration) on doses from 1500 to 3000mg per day. Further kinetic investigations may certainly appear in the near future.

169

Clinical Pharmacokinetics of Ant iconvulsants

3. Carbamazepine (CBZ) Although carbamazepine has been shown to be a primary anticonvulsant of the same degree of effectiveness as phenytoin (Simonsen et aI. , 1975; Troupin et al. , 1975), in Europe it is still used as the drug of third choice (Guelen et aI. , 1975). With an antiepileptic effect similar to phenytoin, and with considerably fewer side-effects, it could become the drug of choice in the treatment of grand mal and psychomotor epilepsy (Grant, 1976), although full agreement on this statement does not exist. It is, however, still a much more expensive drug than phenytoin. The clinical pharmacokinetic profile of carbamazepine is summarised in table II. 3.1 Basic Human Pharmacokinetics

3.1.1 Absorption While rapidly absorbed from a propylene glycol solution, absorption from the clinically used tablets is slow (maximum concentration 6 to 18 hours after single dose) . The fraction absorbed appears to be at least over 70% (Morselli et aI., 1975; Faigle and Feldmann, 1975 ; Levy et aI., 1975a). Possible interindividual variability in the absorption remains to be investigated. Given on an empty stomach the bioavailability is reduced (Levy et aI. , 1975a). 3.1.2 Distribution The apparent volume of distribution is 0.79 to 1.40L/kg (Morselli et aI., 1975; Rawlins et aI., 1975). Plasma protein binding is 69 to 73%. Interindividual variation in binding and the influence of other antiepileptics are small (Johannessen and Strandjord, 1972; Morselli et al., 1975; Hooper et al., 1975). 3.1.3 Elimination The plasma half-life is 30 to 50 hours after a single dose , but shorter after multiple dosing (see section 3.1.4). Carbamazepine is found to an

Table II. Cli nica l pharmaco kin et ic prof il e of carbamazepin e Property

Notes

Absorption rate

t max 6 to 18h (may depend on pharma· ceut ical preparation)

Absorption fraction

70% (may depend on pharmaceutical preparation)

Apparent volume of distribution 0.8 to 1.40L/kg Distribution to CNS, CSF

Probably fast

Protein-binding

60 to 73%

Elimination: biotransformation

'V99% of the absorbed. Appr. 30% excreted with faeces

Elimination : renal (unchanged)

<1 %

Apparent plasma half·life

Single dose 30 to 60h; continuous treatment down to less than 20h (considerable individual variation)

Plasma clearance

'VO.2L/kg/h (van der Kleijn, 1975)

Pattern of elimination kinetics

First order

Kinetic model applicable

Two compartment

Metabolites

1 O-ll-epoxide (and others). Probably act ive

Steady·state/dose ratio

3 to 6-fold variation

Therapeutic plasma range

4 to 10JJg/ml

Dose schedule

2 to 3 to (4) times daily. Pronounced self· induction

Special features

Some interactions

extent of 28% in the faeces, more than half of this as metabolites, and is presumably eliminated via the bile. Only about 2% of the administered dose was recovered in the urine (Faigle and Feldmann,

Clinical Pharmacokinetics of Anticonvulsants

170

1975; Levy et al., 1975a), pointing to an almost complete biotransformation. Of the metabolites excreted through the kidneys, CBZ-I0,1l-epoxide accounts for 1 to 2% and trans-dihydro-dihydroxyCBZ for about 20% (Morselli et aI., 1975; Faigle and Feldmann, 1975; Levy et al., 1975a). The remaining part of the metabolites has not been .accounted for. 3 to 5% of carbamazepine is found in plasma as CBZ-I0,II-epoxide after single dose , approx· imately 15% after multiple doses (Eichelbaum et al., 1975). The epoxide is biologically active in rats (Morselli et al., 1975). It is about 50% bound to plasma proteins (Morselli et aI., 1975), and it has been suggested that this metabolite is partly responsible for the side-effects of carbamazepine (see section 3.3.1).

Specific studies with quantitation of both liver disease and possible defects of the plasma protein are not available. In newborns, the half-life of carbamazepine was found to be in the same range as found in adults after mUltiple doses (Rane et aI., 1975), but the drug had been received transplacentally as the mothers were treated with carbamazepine and other anticonvulsants. It is quite possible that induction of the fetal drug metabolising enzymes had taken place. In children, the steady state plasma levels were in the same range as seen in adults on corresponding doses. The relative concentra tion of the epoxide was found to vary to the same degree as in adults (Rane et aI., 1976).

3.1.4 Elimination Kinetics The plasma kinetics of carbamazepine can best be described by a two compartment open model with fust order kinetics. Thus, serum level increases proportional to the increase in dose (Troupin et al., 1975). Of significance for the clinical use of carb amaze pine , is the change in serum half-life after a few weeks of treatment. A reduction of the half-life from about 35 to about 20 hours is no t rare (Levy et aI., 1975b; Eichelbaum et al., 1975; Strandjord and Johannessen, 1975). This is due to self-induction and means that single dose studies prior to treatment cannot be expected to have predictive value for the steady state level.

3.3 Implications of Clinical Pharmacokinetic Properties to Therapy

3.2 Influence of Disease and Physiological States on Kinetics

Renal and liver disease: The protein binding capacity in plasma is unaffected in patients with renal disease, whereas plasma from patients with hepatic disease binds a slightly lower percentage of CBZ than normal plasma (Hooper et al ., 1975). This is probably without clinical Significance.

3.3.1 Relation of Activity to Plasma Level Therapeutic effect: Very few controlled clinical studies have been done . Previous suggestions of a 'therapeutic plasma level' (4 to 10J.lg/ml) were mostly based on clinical impressions. Troupin et al. (1975) on the basis of some (not very controlled) studies suggested a level between 7 and 16J.l~/ml for the treatment of grand mal, whereas Dam et al. (1975) found a level of above 4J.lg/ml to be adequate. These authors were not able to define any therapeutic level for psychomotor seizures. Schneider (1975) found in institutionalised patients, good seizure control on an average of 4.6 ± 1.3J.lg/ml and in hospitalised patients 6.5 ± 3. OJ.lg/ml , both fasting serum levels. The clinical impressions of a therapeutic range of 4 to 10J.lg/ml thus seem to be correct. However, some disagreement with this point of view may be concluded from other investigations. Thus Eichelbaum et al. (1976) found no further reduction in seizure frequency in patients treated with phenytoin, when carbamazepine was combined a t a plasma level of about 5J.lg/ml. At a carbamazepine level of

Clinical Pharmacokinetics of Anticonvulsants

7 to 8J.1g/ml this combination only caused sideeffects. One explanation could be the accelerated rate of carbamazepine metabolism caused by phenytoin, resulting in a higher concentration of the epoxide, which might be responsible for part of the therapeutic effect and the side-effects. The therapeutic plasma range of carbamazepine may accordingly depend on the concentration of the epoxide, which in turn depends on whether the patient is treated with carbamazepine alone or in combination with phenytoin, phenobarbitone or primidone (Dam et aI. , 1975 ; Morselli et aI., 1976). Side-effects: (i.e. an upper therapeutic limit) is also unclear, as stated above. It is well known that tolerance to side-effects develops in almost all patients. If the dose is gradually increased over 7 to 10 days only a very few patients will experience any side-effects. The maximum intensity of the ' side-effects does not always seem to follow the serum concentration (Levy et aI., 1975a). An early production of a potent metabolite might be able to elicit the side-effects (see below). Meinardi (1972) described side-effects commencing at levels over 2J.1g/ml, with the incidence increasing rapidly until 40 to 50% of the subjects were affected at levels of 8.5 to 10J.lg/ml. Nystagmus was present in an appreciable number of his patients with serum carbamazepine levels of 1.5J.1g/ml. These observations are puzzling and are in contrast to Schneider (1975), who noted the first appearance of side-effects beyond a threshold of 8.9J.1g/mi. It has been suggested that part of the clinical effect and the side-effects of carbamazepine treatment might be caused by the metabolite, CBZ-IO,ll-epoxide (Dam et aI. , 1975), as it was shown that doses of 25 to 50 and 100mg/kg intraperitoneally to rats had a definite protective effect toward maximum electroshock convulsions, which in the case of l00mg/kg lasted up to 6 to 8 hours (Morselli et aI., 1975). The apparent plasma half-life of the epoxide is Significantly shorter than that of carbamazepine. However, the epoxide is only bound to plasma

171

proteins to the extent of about 48 to 53% (Morselli et al ., 1975 ; Eichelbaum et al., 1976), which means that the active free part is greater than the free fraction of carbamazepine. Intoxication seems to occur with lower blood levels of carbamazepine in patients whose basic anticonvulsant blood levels are highest, with symptoms and signs as the blood level approaches values of 4J.1g/ml or over (Kutt et aI., 1975). This might support the theory that part of the intoxication is caused by the metabolite, as the concentration of the epoxide is highest in patients on combined therapy, cf. above (Christiansen and Dam, 1975; Dam et aI., 1975). 3.3.2 Dose Problems Both the preferable dose and dose interval are dependent on the half-life and as this parameter changes from single to multiple dose regimen (see above), dose-schedules are not readily precalculated. A dose regimen of 3 to 4 daily doses might often be necessary. Drug interactions (see below) will complicate this further. 3.3.3 Drug Interactions In patients treated with carbamazepine in combination with phenytOin, phenobarbitone or both the plasma level of carbamazepine is significantly lower than the plasma carbamazepine of patients treated with carbamazepine alone (Christiansen and Dam, 1973; Johannessen and Strandjord, 1975b). Increased metabolism of carbamazepine induced by phenytoin and phenobarbitone was suggested and later proven by an increased plasma CBZ-l 0,11 ,-epoxide during combined treatment (Christiansen and Dam, 1975; Dam et aI., 1975). This interaction is of great clinical relevance, as the plasma level of carbamazepine fluctuates very much in patients on combined medication compared with patients treated with carbamazepine alone. The result is that patients may suffer from seizures in the morning, when plasma carbamazepine is low, and show signs and symptoms of intoxication in the

Clinical Pharmacokinetics of Anticonvulsants

evening, when the plasma level is high. The apparent half-life of CBZ in plasma from patients on combined therapy is shorter than the half-life in patients only treated with carbamazepine (Dam and Christiansen, 1976; Troupin et al., 1974). In patients with seizures in the morning who are being treated with phenobarbitone, phenytoin or primidone, it is recommended to administer carbamazepine in 4 daily doses, with the largest dose given as late as possible in the evening (Dam and Christiansen, 1976). Carbamazepine itself seems to be able to induce liver microsomal enzymes (Morselli et aI., 1972) and thereby also to enhance the elimination of other drugs. This explains the fall in plasma phenytoin level in some patients treated with phenytoin and carbamazepine, and the accelerated metabolism of warfarin (Hansen et aI., 1971). This interaction with phenytoin has not been shown to be of defmite relevance in clinical practice. Carbamazepine significantly reduces the half· life of doxycycline, the only tetracycline to be metabolised (penttilii et aI. , 1974; Neuvonen et aI. , 1975).

3.4 Conclusions Carbamazepine is one of the most important antiepileptic drugs. It is most probably as effective as phenytoin, but with fewer side-effects. Because of its first order kinetics, it is considerably easier to use than phenytoin in anticonvulsive treatment. One reservation however, is that although much is known about the metabolism of carbamazepine, the problem of active metabolites (in particular the epoxide) and their clinical significance is still unsolved. Because of self-induction, changes in the half-life on combined treatment with phenytoin, phenobarbitone or prirnidone and a fairly well established therapeutic plasma range, plasma monitoring is recommended. If possible this should include the epoxide metabolite.

172

4. Phenobarbitone Previously, phenobarbitone was the most important anticonvulsant. Now it plays a more limited role, but it is still a drug of considerable importance. In contrast to its time-honoured place in therapy its kinetic behaviour (table III) in man has not been explored to the same extent as for example phenytoin. Although other barbiturates are used as anticonvulsants, only phenobarbitone will be discussed in this review.

4.1 Basic Human Pharmacokinetics

4.1.1 Absorption Quantitative data are scarce. From the work of Svensmark and Buchthal (1963) it is likely that about 80% is absorbed, but this may vary with the pharmaceutical formulation. The rate of absorption may also vary. Thus, Lous (1954) found peak concentrations in plasma 6 to 18 hours after intake. 4.1.2 Distribution The concentrations of phenobarbitone in the brain, spinal fluid and plasma (unbound) correlated very well in humans (Houghton et aI. , 1975). Protein binding in plasma is about 45 %. The apparent volume of distribution is 0.7 to 1L/kg. pH plays a significant role in the distribution (and the elimination) of phenobarbitone due to its pka-value being close to the physiological pH. A decrease in pH causes a fall in plasma concentration, but the tissue concentration changes in the opposite direction (Waddell and Butler, 1957). 4.1.3 Elimination Phenobarbitone is both metabolised in the liver and excreted unchanged through the kidneys. The distribution between these two main routes of elimination is, however, still uncertain. In re-

Clinical Pharmacokinetics of Anticonvulsants

Table III. Clinical pharmacokinetic profile of pheno· barbitone Property

Notes

Absorption rate

t max 6to 18h m.ln· formation insufficient

Absorption fraction

Up to 80%. Information insufficient

Apparent volume of distribution 0.7 to 1 L/kg Distribution to CNS, CSF

Fast, cone. equilibrium with plasma

Protein-binding

'\45%

Elimination : biotransformation

Approx. 65%. Subject to variation

Elimination: renal (unchanged)

Approx. 35%. Subject to variation

Apparent plasma half-life

50 to 120 to (140)h; children: 40 to 70h

Plasma clearance

'VO.004L/kg/h (van der Kleijn, 1975)

Pattern of elimination kinetics

First order

Kinetic model applicable

Two compartment (7)

Metabol ites

Inactive

Steady·state/dose ratio

Very variable

Therapeutic plasma range

10 to 25 (30) J,lg/ml

Dose Schedule

Once daily , children often twice daily

Special featu res

Pronounced induction of other drugs. Some interactions

viewing the literature, Lous (1966) noted 10 to 30% to be unmetabolised. This is in agreement with Ravn-Jonsen et aI. (1969), who found about two-thirds of phenobarbitone to be excreted as metabolites in poisoned cases. However, an early report (Butler et al., 1954) found about two-thirds unchanged phenobarbitone in the urine. KKllberg

173

et al. (1975) seem to fmd less than 50% in the urine as unchanged phenobarbitone. These apparent discrepancies are, however, very important as they indicate an inter- and maybe also an intra- individual variation as the determining factors in the elimination of phenobarbitone. The pH of the urine is of importance in this connection as alkaline urine facilitates excretion of unchanged phenobarbitone. The half-life in adults varies between 50 to 140h. The ability of phenobarbitone to induce microsomal drug metabolism (in man as well as animals), is well known. Self-induction takes place, but it has never been convincingly demonstrated to be of clinical Significance. Metabolites: The major metabolite is parahydroxy-phenobarbitone (Algeri and McBay, 1956; KKllberg et aI., 1975), most of which is excreted non-conjugated. This metabolite is biologically inactive. Elimination kinetics: First order kinetics have uniformly been found in all studies. 4.2 Influence Kinetics

of

Physiological

States on

Children: The half-life of phenobarbitone varies with age (Jalling 1974), being shorter in children (see section 4.3.1). In the elderly, creatinine clearance decreases (Kampmann et al., 1971), and a decreased urinary output of phenobarbitone should therefore be expected. Plasma half-life or plasma levels have, however, not been investigated in controlled studies in the older age groups, and it is therefore not possible to verify pOSSible kinetic differences (cf. K~berg et al., 1975). It should in this connection be remembered that barbiturates are more likely to cause side-effects in elderly patients. Liver disease: Severely decreased hepatic function should be expected to retard the rate of biotransformation of phenobarbitone, but very few reports are available. Lous (1954) described ele-

174

Clinical Pharmacoki netics of Anticonvulsants

vated serum levels in patients with liver disease. For some other barbiturates it has not been possible to demonstrate a prolongation in the plasma half-life in such patients. Renal disease: Although no specific reports are available, renal impairment must be expected to increase the half-life of phenobarbitone. Increased toxicity in patients with renal disease is observed clinically (Lous, 1966). Acidic urine will increase the reabsorption, thus prolonging the half-life (cf. above).

4.3 Implications of Clinical Pharmacokinetic Properties to Therapy

4.3.1 Relation of Activity to Plasma Level Several controlled studies (cf. Buchthal and Lennox-Buchthal, 1972) have confirmed a thera· peutic plasma range of 10 to 251lg/ml. A level above 30llg/ml is often associated with toxic symptoms. A steady state level will not be reached until after a period of 2 to 4 weeks, because of the long half-life. In children the half-life is shorter (40 to 70h), resulting in a lower plasma level/dose ratio and a shorter period for reaching the therapeutic plasma level. Because of the shorter half· life, higher doses per kg body weight are needed in children. 4.3.2 Drug Interactions Phenobarbitone is a wellknown inducer of hepatic microsomal enzyme drug metabolising activity. Concomitant medication decreases the plasma levels of dicoumarol (bishydroxycoumarin) and warfarin, with a decrease in prothrombin response (Cucinell et al., 1965). A rebound effect may be deleterious if phenobarbitone is stopped and the anticoagulant drug is continued at the same dosage as before. Significant induction of phenytOin by phenobarbitone is not often seen, and has apparently no clinical relevance. More important is the effect on the metabolism of carbamazepine (see above). Concomitant treat-

ment with sodium dipropyl acetate (sodium valproate) causes an increase of the plasma level of phenobarbitone (see section 8.3.2; Schobben et al.,1975).

4.4 Conclusion Phenobarbitone is still a much used anticonvulsant, but its human kinetics have not been very well investigated. The main problems are the variability in the rate of elimination (both metabolism and excretion) and the long plasma half-life . Children need higher doses per kg body weight to obtain optimum plasma levels. A therapeutic plasma range has been established and monitoring of plasma levels is recommended.

5. Primidone Primidone resembles the barbiturates to which it is chemically and pharmacologically related. It has been used for the treatment of grand mal and psychomotor seizures for more than two decades, but it is still a matter of dispute whether the clinical effect of primidone is due to the parent compound as well as to one of the metabolites phenobarbitone , or to the latter alone. Its clinical pharmacokinetic properties are summarised in table IV.

5.1 Basic Human Pharmacokinetics

5.1.1 Absorption Only scarce information is available . Both Booker et al. (1970) and Gallagher and Baumel (1972) found the average t max after single oral doses to about 3 hours with a fairly wide range (0.5 to 9 hours). After continued administration the absorption was slower, but this problem as well as the magnitude of the absorbed fraction do not seem to have been investigated closely.

175

Clinical Pharmacokinetics of Anticonvulsants

Table IV. Clin ical pharmacokinetic profile of primidone Property

Notes

Absorption rate

t max : 0 .5 to 9h

Absorption fraction

Not known

Apparent volume of distribution 'V0 .6L/kg Distribution to CNS, CSF

Fast

Protein-binding

0 (20%?)

Elimination : biotransformation

Totally

Eliminat ion: renal (unchanged)

None

Apparent plasma half-I ife

10 to 12h. Shorter after continuous administration

Plasma clearance

'VO.06L/kg/h (van der Kleijn, 1975)

Pattern of elimination kinetics

First order

Kinetic model applicable Metabolites

PEMA and phenobarbitone . Both active

Steady-state/dose ratio Therapeutic plasma range

5 to 10J.Lg/ml. Problems of phenobarbitone levels must be taken into account

Dose schedules

2 to 3 times daily

5.1.2 Distribution The apparent volume of distribution of primidone itself may be calculated to about 0.6L/kg on the basis of published data. Plasma protein binding is very low «20%) for primidone itself (Houghton et al., 1975) or even negligible (Johannessen and Strandjord, 1974). 5.1.3 Elimination Primidone is almost entirely converted to two metabolites (see below) and almost no primidone

appears unchanged in the urine (Gallagher and Baumel, 1972). The half-life of the parent compound in plasma is about 10 to 12 hours after a single dose, but seems to be shorter after chronic administration. Metabolites: The two main metabolic products of primidone are phenylethyl-malondiamid (PEMA) and phenobarbitone. PEMA is the major metabolite in terms of quantitiy. It is biologically active in animals, although much less so than primidone. Its half-life is (probably) longer and it is present in plasma a few hours after a dose of primidone. Phenobarbitone appears in plasma about 4 days after a continuous medication has been started. As its half-life is considerably longer, it accumulates in plasma (at usual doses of primidone) to a level comparable or even higher than those obtained by giving normal doses of phenobarbitone (Olesen and Dam, 1967; Garrettson, 1972). Elimination kinetics: First order kinetics apply to both main metabolites of primidone. 5.2 Influence of Patho-Physiological States on Kinetics Virtually no information is available. 5.3 Implication of Clinical Pharmacokinetic Properties to Therapy

5.3.1 Relation of Activity to Plasma Level Therapeutic effect: Very incomplete information is available. Furthermore, the conditions are complicated by the fact that phenobarbitone is always present. Booker (1972) found an average plasma level about 7l1g/ml in two matched groups of patients, controlled and uncontrolled for their seizures. Their phenobarbitone and phenytoin levels were also equal. A therapeutic plasma range of 5 to IOl1g/ml has been suggested (Kutt, 1974). Side-effects: It appears that a plasma concentration of 1511g/ml (with therapeutic concen-

176

Clinical Pharmacokinetics of Anticonvulsants

trations of phenobarbitone) is associated with ataxia and/or somnolence (Booker, 1972). A rational basis for plasma level monitoring during prirnidone treatment hardly exists at the present time, but if carried out, it should, of course, include determinations of both primidone and phenobarbitone.

5.3.2 Drug Interactions Few interactions with other drugs have been reported, but what is true for phenobarbitone is probably also true for primidone. The ratio of derived phenobarbitone to unmetabolised primidone in serum is significantly higher in epileptic patients treated with a combination of primidone and phenytoin, than in patients on prirnidone alone (Reynolds, 1975; Reynolds et al. , 1975). An acceleration of prirnidone oxidation is the most reasonable explanation, but inhibition of phenobarbitone metabolism or an impairment of its renal excretion are also possible. As such combined therapy is common in some countries the interaction may be of clinical importance.

5.4 Conclusion Primidone may be an anticonvulsant in its own right, but it also exerts its anticonvulsant properties through the metabolite phenobarbitone. The clinical kinetics of this combination of a possibly active parent compound and an active metabolite has not yet been thoroughly investigated. An optimum plasma level for primidone is suggested, but monitoring must also include determination of phenobarbitone.

6. Ethosuximide

Ethosuximide is one of the most important drugs for the treatment of absence seizures. Compared with many other anticonvulsant drugs, very little is known about its human pharma-

Table V. Clinical pharmacokinetic profile of etho· suximide Property

Notes

Absorpt ion rate Absorption fraction

Probably complete

Apparent volume of distribution 'VO.7L/kg Distribution to CNS, CSF

Parallel to plasma

Protein-binding

None

Elimination: biotransformation

"'80%

Elimination : renal (unchanged)

"'20%

Apparent plasma half·life

Adults 6Oh, children 30h

Plasma clearance

adults "'1.013L/kg/h (van der Kleijn, 1975)

Pattern of elimination kinetics

First order

Kinetic model applicable

Two compartment (7)

Metabolites

Probably not active

Steady-state/dose ratio

3-fold variation

Therapeutic plasma range

40 to 801lg/ml

Dose schedule

1 to 2 times daily

co kinetics (table V) , even though it has been in use for about 15 years.

6.l Basic Human Pharmacokinetics 6.1.1 Absorption Absorption is rather complete with t max 3 to 7 hours after single oral doses. Protein binding in plasma seems to be negligible and the apparent volume of distribution is about O.7L/kg (adults and children). Tissue distribution has not been investigated in man, but in rats concentrations in most tissues (including brain) follow plasma levels.

177

Cl inical Pharmacokinet ics of Ant iconvu lsants

In man, CSF levels have been shown to parallel the plasma levels (Sherwin and Robb, 1972). The elimination of ethosuximide in man is not fully known. It is mostly metabolised, as only 10 to 20% is excreted as unchanged drug (Chang et al., 1972a; van der Kleijn et ai. , 1973). At least three different metabolites have been identified in urine (predominantly as glucuronides), but none seem to possess anticonvulsant activity (Chang et al., 1972b). Their concentrations in plasma are not known. 6.1.2 Basic Kinetic Pattern All evidence points to first order elimination kinetics, with an apparent plasma half-life of about 60 hours for adults and about 30 hours for children 7 to 9 years of age (Chang et al. , 1972a). No hepatic enzyme induction seems to occur (Sherwin and Robb , 1972).

6.2 Influence of Physiological States on Kinetics No information in this area seems to be available , except for the above stated difference between children and adults . As the volumes of distribution are similar it is likely that either the rate of metabolism or the fraction of renal excretion of unchanged ethosuximide decreases with age.

6.3 Implications of Clinical Pharmacokinetic .Properties to Therapy 6.3.1 Relation of Activity to Plasma Level Although only based on a few controlled studies (Sherwin and Robb , 1972; Penry, 1975), an effective plasma level seems to be about 60~g/rnl (range probably from 40 to 80~g/rnl) in patients who respond to ethosuximide treatment. Dose-related side-effects seem to occur with an unacceptable frequency in patients presenting with

plasma levels above (Chang et al., 1972a).

80~g/rnl

in some reports

6.3.2 Dosing Problems There is a fairly good correlation between doses and steady state plasma level, but interindividual variations (3-fold) must be taken into account. This variation or even greater has also been noted in children (van der Kleijn et ai., 1973). Because of the relatively long half-life (2 to 3 days in adults) a once-a-day dose schedule seems reasonable, and has recently been confirmed in clinical practice (Buchanan et ai., 1976). 6.3.3 Drug Interactions Interaction has been suggested with oral contraceptives, but the clinical significance of this remains to be proven. Other interactions have apparently not been reported for ethosuximide.

6.4 Conclusion The clinical pharmacokinetics of ethosuximide have not been examined closely, but an optimum plasma range has been suggested. Routine plasma determinations may be recommended in clinical practice.

Z Diazepam and Clonazepam The benzodiazepines are of great importance in anticonvulsant therapy. Intravenous diazepam is the drug of choice in status epilepticus and other acute convulsive conditions. The information on diazepam below will therefore mostly concern kinetics after a single parenteral dose (table VI). Until recently , oral benzodiazepines had a very limited role as anticonvulsants, but clonazepam is now finding its place in the spectrum of oral antiepileptic drugs. It may also be used intravenously. The clinical pharmacokinetics of this drug will therefore also be discussed (table VI).

178

Clinical Pharmacokinetics of Anticonvulsants

Table VI. Clinical pharmacok inetic prof ile of d iazepam and clonazepam Property

Diazepam (DZP)

Clonazepam (CNP)

(jv)

(po)

Absorption rate

Probably fast and complete

Absorption fract ion Apparent volume of distribution

1 to 2L/kg

2 to 6L/kg

Distribution to CNS, CSF

Probably fast

Probably fast

Protein-binding

'\196%

82%

Elimination : biotransformation

98 to 99%

98 to 99%

Elimination : renal (unchanged)

2%

2%

Apparent plasma half-life

20 to 95h

19 to 60h

Plasma clearance

'VO.03L/kg/h

'VO.OSL/ kg/h (van der Kleijn, 1975)

Pattern of elimination kinetics

First order

F irst order

Kinetic model applicable

Two compartment (some difficulties)

Two compartment

Metabolites

N·desmethyl·DZP (active)

Steady-state/ dose ratio Therapeutic plasma range

7·amino·CNP 7·acetam ino CNP (both i nact ive) 3 to 10 fold variation

> 400 to 500ng/ml

Dose schedule

7.1 Basic Human Pharmacokinetics

7.1.1 Absorption All benzodiazepines seem to be relatively rapidly absorbed from the gut; tmax in plasma being within 2 to 4 hours for clonazepam (Berlin and Dahlstrom, 1975). Absorption from suppositories is usually quite variable (personal observations), but a new rectal application form ('rectoles') which remedies this defect is soon to become available. Information about the fraction absorbed after oral administration of benzodiaze-

(20) 30 to 60 (70) ng/ml 2 to 3 t imes daily

pines is only available for nitrazepam (53 to 94%; Rieder and Wendt, 1973), but is presumably in the same range for clonazepam. No Significant first pass effect has been detected. Intramuscular administration of diazepam to children and adults does not produce as high levels as after oral administration (Hillestad et al ., 1974). Z1.2 Distribution A two compartment open model with fust order rate constants can be used, although the data sometimes do not fit very well (diazepam: Klotz et

Clinical Pharmacokin et ics of Ant iconvulsants

aI., 1975; Andreasen et aI. , 1976; clonazepam: Berlin and Dahlstrom, 1975). Protein binding is about 96% for diazepam and 82% for clonazepam, and is not affected by the concentration (Rieder, 1973). The apparent volume of distribution is about 1L /kg for diazepam and 2 to 6L/kg for clonazepam (Sjo et aI. , 1975b ; Knop et aI. , 1975; Berlin and Dahlstrom, 1975).

Z1.3 Elimination All benzodiazepines are almost totally metabolised. The terminal half-life of diazepam is variable ; 20 to 95 hours (survey: Greenblatt and Shader, 1974). For clonazepam, the half-life is of the order of 36 hours (l9 to 60 hours) (Kaplan et aI. , 1974; Hvidberg and Sjo, 1975 ; Knop et aI. , 1975; Berlin and Dahlstrom, 1975). Metabolites: After intravenous administration of diazepam, only the N-desmethyl-derivative, which is biologically active , can be detected in plasma in limited amounts (Hillestad et aI. , 1974; Andreasen et aI. , 1976). The two other metabolites, 3-0H-diazepam and oxazepam are not measurable in plasma after a single intravenous dose. The contribution of N-desmethyl-diazepam to the acute anticonvulsant effect is not known, but is probably negligible . After repeated doses of diazepam, N-desmethyl-diazepam accumulates in CSF (Hendel, 1975). Whether or not this affects the acute effect of a later diazepam injection has not been investigated. For orally given CNP (long-term treatment) only the 7-amino-CNP and the 7-acetamino-CNP can be measured in plasma (Sjo et aI. , 1975b; Knop et aI. , 1975), but both are considered biologically inactive. Elimination kinetic pattern: First order elimination seems to exist , and has been confumed by the proportional increase in clonazepam levels with increased dose (Sjo et aI., 1975b; Berlin and Dahlstrom, 1975). However, some observations by Klotz et aI. (1975) seem to suggest some degree of saturation for diazepam after multiple dosing.

179

7.2 Influence of Physiological States on Kinetics

Pregnancy: Diazepam readily crosses the placenta (Erkkola and Kanto, 1973), but the pharmacokinetics of diazepam in the mother are not altered. Information on clonazepam is not available. Neonates: Diazepam is metabolised at a slower rate in the very first period of life (Shannon et aI., 1972; Morselli et aI., 1973; Morselli, 1976). The clinical Significance of this phenomenon for the parenteral use of diazepam in neonates has not been investigated. Children: For clonazepam, it seems likely that the elimination rate is faster in children than in adults (Baruzzi et aI., 1976), but specific studies remain to be done . In the elderly. diazepam is supposed to have a more pronounced effect , but whether or not this is due to altered kinetics is not clear, although some observations would suggest that this is so (Klotz et aI., 1975). Reduced hepatic function results in a slower elimination rate of intravenously given diazepam, but also a slower production of the active metabolite (Klotz et aI. , 1975; Andreasen et aI., 1976). The clinical significance of this for the acute anticonvulsant effect is therefore probably negligible. The influence of hepatic disease on clonazepam kinetics has not been investigated. Renal disease does not seem to affect the elimination of benzodiazepines. A transient decrease in renal function was observed in children after intravenous diazepam administration (Guinard et aI., 1975).

7.3 ImpJica tions of Clinical Pharmacokinetic Properties to Therapy

7.3.1 Relation of Activity to Plasma Level Therapeutic effect: The rapid distribution of diazepam after intravenous administration means

Clinical Pharmacoki net ics of Anticonvulsants

that anticonvulsant plasma levels are not maintained for long periods. The relation between plasma level (peak level ?) and anticonvulsant activity has only been closely investigated for some epileptic manifestations. Booker and Celesia (1973) found that a peak level about 500 to 600ng/ rnl diazepam was necessary to suppress interictal discharges in photosensitive subjects. Hillestad et aI. (1974) found a good correlation between plasma levels and other clinical effects after intravenous injection, with plasma levels falling to 400 to 500ng/rnl after 2 hours. Others, e.g. Harder et aI. (1976), found good correlation between changes in continuous reaction times and the level of diazepam after single intravenous dose to normals, with diazepam concentrations of the same magnitude . It is therefore quite reasonable to suggest that the (acute) plasma levels of diazepam must exceed 400 to 500ng/rnl in order to affect several cerebral functions as well as epileptic seizures. Orally given diazepam rarely results in such levels, but administered rectally it may . For oral medication with c1onazepam, the ratio of plasma level to dose may vary 3 to 10 fold (Sjo et aI. , 1975b; Baruzzi et aI., 1976). In very homogenous patient groups with respect to age, this variation seems, however, to be small (Dreifuss et aI., 1975). Only very few controlled studies have been done to evaluate the possible relationship between plasma levels and antiepileptic effect for clonazepam. Dreifuss et aI. (1975) found a steady state level between 20 and 70ng/ml to be effective in absence seizures, which is the same as observed by Baruzzi et aI. (1976). Based on the plasma levels actually obtained, a therapeutic plasma range of 30 to 60ng/rnl was also suggested in other studies (van der Kleijn et aI., 1975; Sjo et aI. , 1975b). Whether this holds true for all types of epileptic seizures has not been investigated. Side-effects: The relation of plasma level to side-effects is also a difficult problem to solve because tolerance develops. Acute side-effects to diazepam have been seen above 200 to 400ng/rnl, but it has not been evaluated if the rare respiratory

180

depression seen after intravenous injection of diazepam coincides with extraordinarily high plasma levels. For c1onazepam, psychic side-effects (dysphoria) may develop if the plasma concentration exceeds 60ng/ rnl for some weeks (Sjo et aI. , 197 5b), but no correlation was found between other side-effects (sedation, dysco-ordination, drowsiness etc.) and plasma levels. The latter was confumed by Baruzzi et aI. (1976) , who also agree that it is not very useful to monitor plasma c1onazepam in order to avoid these side-effects. However, Baruzzi et al. (1976) also found that severe toxic effects developed in the majority of patients with a steady state level above 100ng/m!. Thus, increased frequency of seizures was observed in several of these patients, and levels exceeding 180ng/rnl resulted in status epilepticus. These authors, therefore , recommend monitoring of clonazepam plasma levels during continuous treatment, but more controlled studies should be conducted before this value of monitoring clonazepam can be answered definitively. For clinical use , prediction of steady state plasma levels of clonazepam can be carried out by a pre-dose testing, and clearance can be calculated individually (Knop et aI., 1975; Berlin and Dahlstrom, 1975). In this way the above mentioned serious toxic effects could possibly be avoided. Z3.2 Drug Interactions Only a few and probably clinically inSignificant kinetic interactions have been described for diazepam. Reduction in the steady state level of clonazepam by a probable induction phenomenon caused by phenytoin and/or phenobarbitone has been noted (Sjo et aI., 1975b; Baruzzi et aI. , 1976). The clinical significance of this is, however, still not evaluated. The addition of clonazepam to treatment with phenytoin and other anticonvulsants has given conflicting results about the influence of clonazepam on their steady state levels (cf. Baruzzi et aI. , 1976).

181

Clin ical Pharmacokinetics of Ant iconvu lsants

7.4 Conclusion For anticonvulsant treatment with the benzodiazepines plasma level monitoring is generally not advocated. For long-term treatment with clonazepam orally, plasma level monitoring might, however, be advantageous, or even necessary in cases with suspicion of toxic effects (e.g. increased seizure frequency , severe dysphoria etc.) or if interaction (e .g. with phenytoin or phenobarbitone) is a possibility.

8. Di-N-Propylacetate (Sodium Valproate) The sodium salt of di-n-propylacetic acid was introduced as an antiepileptic drug in 1967. The chemical structure, which has been known for almost 100 years , is different from all other antiepileptic drugs. In clinical trials it has appeared to be effective against absences in petit mal epilepsy, and in combination with phenobarbitone in the treatment of grand mal epilepsy. Essentially only two human pharmacokinetic studies are available at the present time . 8.1 Basic Human Pharmacokinetics

8.1 .1 Absorption After oral administration t max in plasma is within 1 h our (Eymard et aI. , 1972, cited fro m Schobben et aI. , 1975). Milk products did not significantly influence the maximal plasma concentration, nor the time at which the peak was reached (Schobben et al. , 1975). A more gradual absorption would be desirable as the elimination is quite rapid. A preliminary approach to slow down the rate of absorption has been done by Sonnen et al. (l975). 8.1.2 Distribution The relative apparent distribution volume, 0.15 to 0.40Ljkg, is small in comparison with other antiepileptic drugs.

8.1.3 Elimination Sodium valproate is probably totally metabolised (as in the rat and mouse) as the renal excretion (in 1 patient) was small (Schobben et aI. , 1975). In 6 epileptic children, aged 3 to 12 years, treated with sodium valproate for 10 to 20 days in combination with phenobarbitone, phenytoin, ethosuximide or clonazepam the half-life was found to be between 8.0 and 11.7 hours. Elimination kinetics: In the patients described above the plasma concentration declined monoexponentially. The correlation between dose and plasma concentration was rather poor (Schobben et al., 1975). 8.2 Influence of Patho-physiological States on Kinetics This aspect has not been investigated. 8.3 Implications of Clinical Pharmacokinetic Properties to Therapy

8.3.1 Relation of Activity to Plasma Level Therapeutic effect: An optimum plasma level has not been demonstrated as most of the pa tients investigated were treated with combinations of other antiepileptic drugs. The usual level seems to be 60- 80/-Lgjrnl (Sonnen et al. , 1975). Side-effects: The relationship between plasma concent ration and side-effects has not been studied. Sodium valproate is well tolerated. Only gastric symptoms have been observed in some patients. 8.3.2 Drug Interactions A hypnotic effect is probably caused by a rise in plasma phenobarbitone in patients on combined medication (Schobben et aI. , 1975). If phenobarbitone is given along with sodium valproate the dose of phenobarbitone should be reduced. Kinetic interaction with phenytoin was not observed. Concomitant treatment with levodopa resulted in a 50% increase of the half-life

Clinical Pharmacokinetics of Anticonvulsants

(Schobben et al., 1975). The mechanism of these possible drug interactions is not clear.

8.4 Conclusions This interesting new drug remains to be explored kinetically in patients. It is too early to state to which extent pharmacokinetic measurements will benefit the use of sodium valproate.

Acknowledgement We are very grateful for discussion and technical comment from Drs Mogens Lund and Per Buch Andreasen. This study was partly supported by grants from The Danish State Medical Research Council.

References Affrime, M. and Reidenberg, M.M.: The protein binding of some drugs in plasma from patients with alcoholic liver disease. European Journal of Oinical Pt.armacology 8: 267-269 (1975). Albert, K.S.; Hallmark, M.R.; Sakmar, E.; Weidler, DJ. and Wagner, J.G. : Plasma concentrations of diphenylhydantoin, its parahydroxylated metabolite, and corresponding glucuronide in man. Research Communications in Chemical Pathology and Pharmacology 9: 463-469 (1974). Algeri, RJ . and McBay, A.J.: Metabolite of phenobarbital in human urine. Scienoe 123: 183-184 (1956). Alvan, G.; Bertler, A.; Eef-Olofsson, 0.; Karlsson, E.; Sjoqvist, F. and Tomson, G.: Biological availability A comparison of three phenytoin preparations. Liikartidningen 72: 2621-2623 (1975). Andreasen, Frederik: Protein binding of drugs in plasma from patients with acute renal failure. Acta pharmacologica et toxicologica 32: 417-429 (1973). Andreasen, P.B.; Fr~nd, A.; Skovsted, L.; Andersen, S.Aa. and Hauge, M.: Diphenylhydantoin half-life in man and its inhibition by phenylbutazone: The role of genetic factors. Acta Medica Scandinavica 193: 561-564 (1973). Andreasen, P.B.; Hendel, J. ; Greisen, G. and Hvidberg, E.F.: Pharmacokinetics of diazepam and quantitative

182

liver function. European Journal of Oinical Pharmacology (In press, 1976). Arnold, K. and Gerber, N.: The rate of decline of diphenylhydantoin in human plasma. ainical Pharmacology and Therapeutics 11 : 121-134 (1970). Atkinson, A.J. and Shaw, J.M. : Pharmacokinetic study of a patient with diphenylhydantoin toxicity. ainical Pharmacology and Therapeutics 14 : 521-528 (1973). Bach, B. ; Kampmann, J.P.; Rasmussen, S.N.; Skovsted, L. and Molholm Hansen, 1. : Correlation between liver volume and the rate of metabolism of phenytoin and antipyrine. In preparation. Barth, N. ; Alvan, G.; Borg!i, O. and Sjoqvist, F.: Two-fold interindividual variation in plasma protein binding of diphenylhydantoin and warfarin in patients with epilepsy. Archives of Neurology (In press, 1976). Baruzzi, A.; Gerna, M.; Altamura, A.C.; Avanzini, G.; Canger, R.; Zagoni, P. and Morselli, P.L. : aonazepam plasma levels in epileptic patients in course of chronic treatment. Epilepsia (In press, 1976). Berlet, H. : Serum levels of phenytoin in children; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) ainical Pharmacology of Anti-Epileptic Drugs, p.63-69 (Springer-Verlag, Berlin 1975). Berlin, A. and Dahlstrom, H.: Pharmacokinetics of the anticonvulsant drug clonazepam evaluated from single oral and intravenous doses and by repeated oral administration. European Journal of Oinical Pharmacology 9: 155-159 (1975). Blaschke, T.F.; MefflIl, P.J .; Melmon, K.L. and Rowland, M. : Influence of acute viral hepatitis on phenytoin kinetics and protein binding. Clinical Pharmacology and Therapeutics 17 : 685 - 691 (1975). Brennan, R.W.; Dehejia, H.; Kutt, H.; Verebeley, K. and McDowell, F.: Diphenylhydantoin intoxication attendant to slow inactivation of isoniazid. Neurology 20: 687-693 (1970). Bochner, F.; Hooper, W.D. ; Sutherland, J.M.; Eadie, M.J. and Tyrer, J.H.: Diphenylhydantoin concentrations in saliva. Archives of Neurology 31 : 57-59 (1974) . Booker, H.E. : Primidone. Relation of plasma levels to clinical control; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.373 - 376 (Raven Press, New York; North-Holland, Amsterdam 1972). Booker, H.E.; and Celesia, G.G. : Serum concentrations of diazepam in subjects with epilepsy. Archives of Neurology 29: 191-194 (1973). Booker, H.E. ; Hosokowa, K.; Burdette, R.D. and Darcey, B. : A clinical study of serum primidone levels. Epilepsia 11 : 395-402 (1970). Borondy, P.; Dill, W.A.; Chang, T.; Buchanan, R.A. and Glazko, A.1. : Effect of protein binding on the distribution of 5,5-d.iphenylhydantoin between plasma and

Clin ical Pharmacokinetics of Anticonvulsants

red cells; in Anton and Solomon (Eds) Drug-Protein Binding, p.82- 87 (New York Academy of Sciences 1973). Boston Collaborative Drug Surveillance Program: Diphenylhydantoin side effects and serum albumin levels. Oinical Pharmacology and Therapeutics 14 : 529 - 532 (1973). Buchanan, R.A ; Kinkel, A.W.; Goulet, 1.R. and Smith, T.e.: The metabolism of diphenylhydantoin (Dilantin ®) following once-daily administration . Neurology 22 : 126 - 130 (1972) . Buchanan, R.A ; Kinkel, A.W.; Turner, 1.L. and Heffelfmger, I.e. : Ethosuximide dosage regimens. ainical Pharmacology and Therapeutics 19 : 143 - 147 (1976). Buchthal, F. and Lennox-Buchthal, M.A.: Phenobarbital. Relation of serum concentration to control of seizures; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.335-343 (Raven Press, New York; North-Holland, Amsterdam 1972). Buchthal, F. and Svensmark, 0. : Aspects of the pharmacology of phenytoin (Dilantin ®) and phenobarbital relevant to their dosage in the treatment of epilepsy. Epilepsia 1: 373-384 (1960). Butler, T.e. ; Makaffee, C. and Waddell, W.l .: Phenobarbital : Studies of elimination, accumulation, tolerance and dosage schedules. 10urnal of Pharmacology and Experimental Therapeutics Ill: 425-435 (1954). Cereghino, 1.1.; Brock, I.T. and Penry, 1.K.: Albutoin; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.283-291 (Raven Press, New York; NorthHolland, Amsterdam 1972). Chang, T.; Dill, W.A and Glazko, AI. : Ethosuximide. Absorption, distn'bution and excretion; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.417 - 423 (Raven Press, New York; North-Holland, Amsterdam 19718). Chang, T.; Burkett, A.R.and Glazko, A.J.: Ethosuximide. Biotransformation; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.425- 429 (Raven Press, New York; North-Holland, Amsterdam 1972b). Christensen, L.K. and Skovsted, L : Inhibition of drug metabolism by chloramphenicol. Lancet 2: 1397 - 1399 (1969). Christiansen, J. and Dam, M. : Influence of phenobarbital and diphenylhydantoin on plasma carbamazepine levels in patients with epilepsy. Acta Neurologica Scandinavica 49: 543-546 (1973). Christiansen, 1. and Dam, M.: Drug interaction in epileptic patients; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.197 -199 (Springer-Verlag, Berlin 1975).

183

Christiansen, e. ; Naestoft, 1. ; Hvidberg, E.F.; Larsen, N.-E. and Petersen, B. : An easy procedure for in vivo estimation of protein binding and correction of elevated serum values induced by venous stasis. Clinica Chimica Acta 62 : 65 - 71 (1975). Cocks, D.A. ; Critchley, E.M.R.; Hayward, H.W.; Owen, V.; Mawer, G.E. and Woodcock, B.G. : Control of epilepsy with a single daily dose of phenytoin sodium. British Journal of Oinical Pharmacology 2: 449 - 453 (1975). Cook, e.E. ; Amerson, E.; Poole, W.K.; Lesser, P. and O'Tuama, L : Phenytoin and phenobarbital concentrations in saliva and plasma measured by radioimmunoassay. Clinical Pharmacology and Therapeutics 18: 742- 747 (1975). Cucinell, S.A. ; Conney, A.H. ; Sansur, M.S. and Burns, 1.1. : Drug interactions in man. Lowering effect of phenobarbital on plasma levels of bishydroxycoumarin (Dicoumarol ®) and diphenylhydantoin (Dilantin ®). Oinical Pharmacology and Therapeutics 6 : 420-429 (1965) . Dam, M.; Hiibbe, P. and Christiansen, 1. : Etnisk forskel i serum-fenytoin. Ugeskrift for Laeger 136 : 1974-1976 (1974). Darn, M.; lensen, A and Christiansen, 1.: Plasma level and effect of carbamazepine in grand mal and psychomotor epilepsy. Acta Neurologica Scandinavica Suppl. 60: 33-38 (1975). Dam, M. and Christiansen, 1.: Carbamazepine: The clinical effect of drug interaction. Proceedings from Epilepsia-75, Helsinki (In press 1976). Dam, M.; Mygind, K.J. and Christiansen, J.: Antiepileptic drugs; Plasma clearance during pregnancy. Proceedings from Epilepsia-75 Helsinki (In press 1976). Dam, M. and Olesen, O.V.: Intramuscular administration of phenytoin. Neurology -(Minneap.) 16 : 288- 292 (1966). Dreifuss, F.E.; Penry, I.K.; Rose, S.W.; Kupferberg, H .J. ; Dyken, P. and Sato , S. : Serum clonazepam concentrations in children with absence seizures. Neurology 25 : 255- 258 (1975). Eadie, M 1'.: Plasma level monitoring of anticonvulsants. Clinical Pharmacokinetics I : 52-66 (1976). Ehrnebo, M. ; Agurell, S.; Jailing, B. and Borc~us , L.O.: Age differences in drug binding by plasma proteins : Studies on human foetuses, neonates and adults. European 10urnal of ainical Pharmacology 3: 189-193 (1971) . Ehrnebo, M. and Odar-Cederlof, I. : Binding of amobarbital, pentobarbital and diphenylhydantoin to blood cells and plasma proteins in healthy volunteers and uraemic patients. European Journal of Clinical Pharmacology 8: 445 - 453 (1975).

Clinical Pharmacokinetics of Anticonvulsants

Eichelbaum, M. ; Ekbom, K. ; Bertilsson, L ; Ringberger, V.A. and Rane, A.: Plasma kinetics of carbamazepine and its epoxide metabolite in man after single and multiple doses. European Journal of Oinical Pharmacology 8: 337-341 (1975). Eichelbaum, M. ; Bertilsson, L ; Lund, 1. ; Palmer, 1. and Sjoqvist, F. : Plasma levels of carbamazepine and carbamazepine-l 0, ll-epoxide during treatment of epilepsy. European Journal of Oinical Pharmacology (In press, 1976). Erkkola, R. and Kanto, J.: Concentration levels of diazepam and its main metabolite N-desmethyldiazepam in the circulatory system of mother and child during delivery; in Bossart, Cruz and Huber (Eds) Perinatal Medicine, p.310-311 (Huber, Berne 1973). Faigle, J.W. and Feldman, K.F. : Pharmacokinetic data of carbamazepine and its major metabolites in man; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) ainical Pharmacology of Anti-Epileptic Drugs, p.159- 165 (Springer-Verlag, Berlin 1975). Gallagher, B.B. and Baumel, J.P.: Primidone. Absorption, distribution and excretion; in Woodbury, Penry and Shcmidt (Eds) Antiepileptic Drugs, p.357 - 359 (Raven Press, New York; North-Holland Amsterdam 1972). Garrettson, LK. : Pharmacology of anticonvulsants. Pediatric Clinics of North America 19: 179- 191 (1972). Glazko, A.l: Diphenylhydantoin metabolism. A prospective review. Drug Metabolism and Disposition 1: 711-714 (1973). Glazko, A.l and Chang, T.: Diphenylhydantoin. Absorption, distribution and excretion; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.130 (Raven Press, New York; North-Holland, Amsterdam 1972). Grant, R.H.E.: Carbamazepine in the treatment of severe epilepsy ; in Birkmayer (Ed) Epileptic Seizures Behaviour - Pain, p.104- 110 (Hans Huber, Bern 1976). Greenblatt, D.l and Shader, R.l.: Benzodiazepines in Clinical Practice (Raven Press, New York 1974). Guelen, P.J.M.; van der Klein, E. and Woudstra, U.: Statistical analysis of pharmacokinetic parameters in epileptic patients chronically treated with antiepileptic drugs; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) ainical Pharmacology of AntiEpileptic Drugs, p.2-10 (Springer-Verlag, Berlin 1975). Gugler, R.; Azarnoff, O:t: and Shoeman, D.W.: Diphenylhydantoin: Correlation between protein binding and albumin concentration. Klinische Wochenschrift 53: 445-446 (1975).

184

Guignard, lP.; Fillonx, B. ; Lavoie, J.; Pelet, J. and Torrado, A. : Effect of intravenous diazepam on renal function. ainical Pharmacology and Therapeutics 18: 401 - 404 (1975). Hansen, J.M.; Kirstensen, M. and Skovsted, 1.: Sulthiame (Ospolot ®) as inhibitor of diphenylhydantoin metabolism. Epilepsia 9: 17-22 (1968). Hansen, I.m.; Kristensen, M.; Skovsted, 1. and Christensen, 1.K.: Dicoumarol-induced diphenylhydantoin intoxication. Lancet 2: 265-266 (1966). Hansen, 1M.; Siersbaek-Nielsen, K.; Kristensen, M.; Skovsted, L and Christensen, 1.K. : Effect of diphenylhydantoin on the metabolism of dicoumarol in man. Acta Medica Scandinavica 189 : 15- 19 (1971). Hansen, J.M.; Siersbaek-Nielsen, K. and Skovsted, 1.: Carbamazepine-induced acceleration of diphenylhydantoin and war farin metabolism in man. Clinical Pharmacology and Therapeutics 12: 539- 543 (1971). Harder, F.; Elsass, P. ; Hendel, J.; Hvidberg, E.F. and Hj~rting-Hansen, E.: ainical and psychological effects of intravenous diazepam related to plasma levels. International Journal of Oral Surgery (In press, 1976). Hayes, M.J.; Langman, M.J.S. and Short, A.H. : Changes in drug metabolism with increasing age: 2. Phenytoin clearance and protein binding. British Journal of ainical Pharmacology 2: 73 - 79 (1975). Hendel, J. : Cumulation in cerebro-spinal fluid of the N-desmethyl metabolite after long term treatment with diazepam in man. Acta pharmacologica et toxicologica 38: 17- 22 (l975}. Hillestad, 1.; Hansen, T.; Melsom, H. and Drivenes, A.: Diazepam metabolism in normal man. l. Serum concentrations and clinical effects after intravenous, intramuscular, and oral administration. Clinical Pharmacology and Therapeutics 16: 479-484 (1974). Hooper, W.D.; Bochner, F.; Eadie, M.J. and Tyrer, lH.: Plasma protein binding of diphenylhydantoin. Effects of sex hormones, renal and hepatic disease. ainical Pharmacology and Therapeutics 15 : 276-282 (1974a). Hooper, W.D. ; Eadie, M.l and Tyrer, J.H. : Plasma diphenylhydantoin levels in Australian children. Australian New Zealand Journal of Medicine 4 : 456-461 (1974b). Hooper, W.D.; Tyrer, J.H. and Eadie, M.J. : Plasma diphenylhydantoin levels in Australian adults. Australian New Zealand Journal of Medicine 4 : 449-455 (1974c). Hooper, W.O.; Oubetz, O.K.; Bochner, P.; Cotter, 1.M.; Smith, G.A.; Eadie, M.J. and Tyrer, lH. : Plasma protein binding of carbamazepine. Clinical Pharmacology and Therapeutics 17: 433-440 (1975).

Clinical Pharmacokinetics of Anticonvulsants

Hoppel, C.; Rane, A. and Sjoqvist, F.: Kinetics of phenytoin and carbamazepine in the newborn ; in Morselli, Garratini and Sereni (Eds) Basic and Therapeutic A,peets of Perinatal Pharmacology, p.341 - 345 (Raven Press, New York 1975). Houghton, G.W. and Richens, A.: Rate of elimination of tracer doses of phenytoin at dif fer ent steady-state serum phenytoin co ncentrations in epileptic patients. British Journal of Oinical Phar macology I: 155- 161 Houghton, G.W. ; Richens, A. and Leighton, M.: Effect of age, height, weight and sex on serum phenytoin concentration in epileptic patients. British Jou rnal of Clinical Pharmacology 2: 251 - 256 (1975 a). Houghton, G.W.; Richens, A.; Toseland, P.A.; Davidson, S. and Falconer, M.A. : Brain concentrations of phenytoin, phenobarbital and primidone in epileptic patients. European Journal of Clinical Pharmacology 9: 73 - 78 (1975b). Hvidberg, E.F. and Sjo, 0 .: Clinical phar maco kinetic experience with clonazepam ; in Schneider , Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharma cology of Anti-Epileptic Drugs, p.242-246 (Springer-Verlag, Berlin 1975). Jalling, B.: Plasma and cerebrospinal fluid concentration of phenobarbital in infants given single dose. Developmental Medicine and Children Neurology 16: 781 - 793 (1974). Johannessen, S.l. and Strandjord, R.E. : The concentration of carbamazepine (Tegretol ®) in serum and cerebrospinal fluid in patients with epilepsy. Acta Neurologica Scandinavica 48 (Suppl. 51): 445-446 (1972). Joh annessen, S.l. and Strandjord, R.E. : Absorpsjon og protein binding i serum av noen antiepileptika. Svenska Epilepsia 4 (Suppl. 2) : 62-67 (1974). Johannessen, S.l. and Strandjord, R.E.: Absorption and protein binding in serum of several anti-epileptic drugs ; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Oinical Pharmacology of AntiEpileptic Drugs, p.262- 273 (Springer-Verlag, Berlin 1975a). Johannessen, S.l. and Strandjord, R.E.: The influence of phenobarbitone and phenytoin on carbamazepine serum levels ; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p. 201 - 205 (Springer-Verlag, Berlin 1975b). Kampmann, J.P.; Siersbaek-Nielsen, K.; Kristensen, M. and M
185

Kaplan, S.A.; Alexander, K. ; Jack, M.L.; Puglise, C.V.; de Silva, J.A.F.; Lee, T.L. and Weinfeld, R.E.: Pharmaco kinetic profiles of clonazepam in dog and humans and of flunitrazepam in dog. Journal of Pharmaceutical Sciences 63: 527-532 (1974). Kleijn, E. van der; Collste, P.; Norlander, B.; Agurel\, S. and Sjoqvist, F.: Gas chromatographic determination of ethosuximide and phensuximide in plasma and urine of man. Journal of Pharmacy and Pharmacology 25: 324-327 (1973). Kleijn, E. van der; Guelen, P.J.M. ; Wijk, e. van and Baars, I. : Oinical pharmacokinetics in monitoring chronic medication with antiepileptic drugs; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.l1 - 33 (Springer-Verlag, Berlin 1975). Klotz, U.; Avant, G.R.; Hoyumpa, A ;'Schenker, S. and Wilkinson, G.R.: The effects of age and liver disease on the disposition and elimination of diazepam in adult man. Journal of Clinical Investigation 55: 347-359 (1975). Knop, H.J.; Kleijn, E. van der and Edmunds, L.C.: The determination of clonazepam in plasma by gasliquid chromatography. Pharmaceutisch Weekblad 110: 297-310 (1975). Kostenbauder, H.B.; Rapp, R.P.; McGovren, J.P.; Foster, T.S.; Perrier, D.G. ; Blacker, H.M.; Hulon, W.e. and Kinkel, A.W.: Bioavailability and single-dose pharmacokinetics of intramuscular phenytoin. Clinical Pharmacology and Therapeutics 18: 449-456 (1975). Kurata, D. and Wilkinson, G.R. : Erythrocyte uptake and plasma binding of diphenylhydantoin. Clinical Pharmacology and Therapeutics 16: 355-362 (1974). Kutt, H.: Pharmacodynamic and pharmacokinetic measurements of antiepileptic drugs. Clinicial Pharmacology and Therapeutics 16: 243-250 (1974). Kutt, H.: Interactions of antiepileptic drugs. Epilepsia 16: 393 (1975). Kutt, H.; Brennan, R.; Dehejia, H. and Verebeley, K.: Diphenylhydantoin intoxication. A complication of isoniazid therapy. American Review of Respiratory Diseases 101 : 377 - 384 (1970). Kutt, H.; Winters, W.; Kohenge, R. and McDowell, F.: Diphenylhydantoin metabolism, blood levels and toxIcity. Archives of Neurology 11: 642-648 Kutt, H.; Winters, W. ; Scherman, R. and McDowell, F. : Diphenylhydantoin and phenobarbital toxicity. Archives of Neurology 11: 649-656 (1964b). Kutt, H.; Wolk, M.; Scherman, R. and McDowell, F. : Insufficient parahydroxylation as a cause of diphenylhydantoin toxicity. Neurology 14: 542-548 (1964c). Kutt, H. ; Solomon, G.; Wasterlain, c.; Peterson, H.; Louis, S. and Carruthers, R.: Carbamazepine in difficult to

Clinical Pharmacokinetics of Anticonvulsants

control epileptic out-patients. Acta Neurologica Scandinavica Suppl. 60 : 27 - 32 (1 975). Kauberg, N.; Agurell, S.; Ericsson, 0. ; Bucht, E. ; Jailing, B. and Boreus, LO. : Quantitation of phenobarbital and its main metabolites in human urine. European Journal of Oinicial Pharmacology 9 : 161-168 (1975). Larsen, N.-E. and Naestoft, J.: Quantitative determination of ethotoin in serum by gas chromatography. Journal of Chromatography 92 : 157- 161 (1974). Levy, R.H. ; Pitlick, W.H.; Troupin, M.D. ; Green, J.R. and Neal, lM.: Pharmacokinetics of carbamazepine in normal man. Cinical Pharmacology and Therapeutics 17 : 657 - 668 (1975a). Levy, R.H.; Pitlick, W.H. ; Troupin, A.S. and Green, lR.: Phar ma co kinetic evaluation of carbamazepine in normal human volunteers following single doses and during chronic administration. Abstracts from the Sixth International Congress of Pharmacology, Helsinki (l975b). Lous, P.: Plasma levels and urinary excretion of three barbituric acids after oral administration to man. Acta pharmacologica et toxicologica 10: 147- 165 (1954). Lous, P.: Elimination of barbiturates; in Johansen, S.H. (Ed) Barbiturate Poisoning and Tetanus, p.3 41-350 ( Little Brown, Boston 1966). Lumholtz, B.; Siersbaek-Nielsen, K. ; Skovsted, L ; Kampman, J.P. and M4lholm-Hansen, J.: Sulfamethizole-induced inhibition of diphenylhydantoin, tolbutamide, and warfarin metabolism. Clinical Pharmacology and Therapeutics 17: 731 - 734 (1975). Lund, L : Effects of phenytoin in patients with epilepsy in relation to its concentration in plasma; in Davies and Prichard (Eds) Biological Effects of Drugs in Relation to their Plasma Concentration, p.227-238 (MacMillan, London 1973). lllnd, L : Clinical significance of generic inequivalence of three different pharmaceutical preparations of phenytoin. European Journal of Oinical Pharmacology 7: 119- 124 (1 974a). Lund, L: Anticonvulsant effect of diphenylhydantoin relative to plasma levels. Archives of Neurology 31 : 289-294 (1974b). Lund, L : Aprospective study of the relationship between plasma levels and effects of phenytoin in grand mal epilepsy. Acta Neurologica Scandinavica Suppl. 60 : 74 (1975). Lund, L and Alvan, G.: Phenytoin dosage nomogram. Lancet 2: 1305 (1975). Lund, L ; Berlin, A. and Lunde, P.K.M. : Plasma protein binding of diphenylhydantoin in patients with epilepsy. Clinical Pharmacology and Therapeutics 13: 196-200 (1972).

186

Lunde, P.K.M. ; Rane, A.; Yaffe, S.l ; Lund , L and Sjoqvist, F.: Plasma protein binding of di phenylhydantoin in man. Qinical Pharmacology and Therapeutics 11 : 846- 855 (1970). Meinardi, H.: Carbamazepine; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.487-496 (Raven Press, New York ; North-Holland, Amsterdam 1972). Mellk, H.M.; Letteri, J.M.; Durante, P.; Louis, S.; Kutt, H. and Glazko, A. : Diphenylhydantoin metabolism i n chronic uremia. Annual Internal Medicine 72 : 801 (1970). Mirkin, B.L : Diphenylhydantoin: Placental transport, fetal localization, neonatal metabolism, and possible teratogenic effects. The Journal of Pediatrics 78 : 329 - 337 (1971). Miyamoto, K. ; Seino, M. and Ikeda, Y.: Consecutive determination of the levels of twelve anti-epileptic drugs in blood and cerebrospinal fluid ; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.323- 330 (Springer-Verlag, Berlin 1975). Morselli, P. L : Clinical pharmaco kinetics in neonates. Clinical Pharmacokinetics 1: 81 - 98 (1976). Morselli, P.L ;Biandrate, P.; Frigerio, A. and Garattini, S.: Pharmacokinetics of carbamazepine in rats and humans. European Journal of Qinical Investigation 2 : 297 (1972). Morselli, P.L ; Bossi, Land Gerna, M.: Pharmacokinetic studies with carbamazepine in epileptic patients ; in Birkmayer (Ed) Epileptic Seizures - Behaviour - Pain p.141 - 150 (Hans Huber , Bern 1976). Morselli, P. L ; Gerna , M.; Maio, G. de; landa, G.; Viani, F. and Garattini, S.: Pharmacokinetic studies on carbamazepine in volunteers and in epileptic patients; in Schneider, Janz, Gardner-Thrope, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.166- 180 (Springer-Verlag, Berlin 1975). Morselli, P.L ; Principi, N.; Tognoni, G. ; Reali, E.; Belvedere, G. ; Standen, S.M. and Sereni, F. : Diazepam elimination in premature and full-term infants and children. Journal of Perinatal Medicine 1: 133 - 141 (1973). N:uvonen, P.J.; Penttila, 0 .; Lehtovaara, R. and Aho, K.: Effect of antiepileptic drugs on the elimination of various tetracycline derivatives. European Journal of Clinical Pharmacology 9: 147- 154 (1975). Naestoft, J.; Hvidberg, E.F. and Sjo, 0 .: Saturable metabolic pathways for ethotoin in man. Clinical and Experimental Pharmacology and Physiology (In press, 1976). Odar-Cederlof, I.: Studies on the plasma protein binding of drugs in patients with renal failure. Thesis, Karolinska Institute, Stockholm (1975).

Clinical Pharmacokinetics of Anticonvu lsants

Odar-CederlOf, I.; Lunde, P. and Sjoqvist, F.: Abnormal pharmacokinetics of phenytoin in a patient with uraemia. Lancet 17: 831-832 (1970). Odar-CederlOf, l. and Borga, 0. : Kinetics of diphenylhydantoin in uraemic patients: Consequences of decreased plasma protein binding. European Journal of Clinical Pharmacology 7: 31-37 (1974). Oelkers, W.; Schafer, H. and Reith, H.: Zur Anwendung eines Phenytoin-Infusionskonzentrates im Kindesalter. Deutsche Medizinische Wochenschrift 100: 1012- 1014 (1975). Olesen, O.V.: The influence of disulfiram and calcium carbimide on the serum diphenylhydantoin. Archives of Neurology 16 : 642-644 (1967). Olesen, O.V. and Dam, M.: Fenytoinindholdet i serum efter intramuskulaer injektion i relation til behandling af status epilepticus. Ugeskrift for Laeger 127: 992- 993 (1965). Olesen, O. V. and Dam, M. : The metabolic conversion of primidone (Mysoline ®) to phenobarbital in patients under long-term treatment. Acta Neurologica Scandinavica 43 : 348-356 (1967). Olsen, G.D.; Bennett, W.M. and Porter, G.A.: Morphine and phenytoin binding to plasma proteins in renal and hepatic failure. Clinical Pharmacology and Therapeutics 17: 677-684 (1975). Penry, J.K. : Correlation of serum ethosuximide levels with clinical effect; in Schneider, Janz, GardnerThorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.217 - 220 (SpringerVerlag, Berlin 1975). Pentikiiinen, P.l.; Neuvonen, PJ. and Elfving, S.M. : Bioavailability of four brands of phenytoin tablets. European Journal of Clinical Pharmacology 9: 213 -218 (1975). Penttila, 0.; Neuvonen, P.J.; Aho, K. and Lehtovaara, R. : Interaction between doxycycline and some antiepileptic drugs. British Medical Journal 2: 470-472 (1974 ). Phillips, P.J.; Pain, R. W. and Duncan, B. McL. : Serumphenytoin levels. Lancet 2: 926 (1975). Pruitt, A.Q. ; Zwiren, G.T.; Patterson, lH. ; Dayton, P.G.; Cook, C.E. and Wail, M.E.: A complex pattern of disposition of phenytoin in severe intoxication. Clinical Pharmacology and Therapeutics 18: 112- 120 (1975). Rane, A.: Urinary excretion of diphenylhydantoin metabolites in newborn infants. Journal of Pediatrics 85: 543 - 545 (1974). Rane, A. ; Bertilsson, L. and Palmer, L.: Disposition of placentally transferred carbamazepine (Tegretol) in the newborn. European Journal of Clinical Pharmacology 8: 283-284 (1975).

187

Rane, A.; Hojer, B. and Wilson, IT.: Kinetics of carbamazepine and its 10-11 epoxide metabolite in children. Clinical Pharmacology and Therapeutics 19: 276-283 (1976). Rasmussen, F.: Excretion of drugs by milk; in Brodie and Gillette (Eds) Handbook of Experimental Pharmacology, New Series Vol. XXVIII / I, p.390-402 (Springer-Verlag, Berlin 1971). Ravn-Jonsen, A.; Lunding, M. and Secher, 0.: Excretion of phenobarbitone in urine after intake of large doses. Acta Pharmacologica Suedica 27: 193-201 (1969). Rawlins, M.D.; Coilste, P.; Bertilsson, L. and Palmer, L.: Distribution and elimination kinetics of carbamazepine in man. European Journal of Clinical Pharmacology 8: 91-96 (1975). Reidenberg, M.M.; Odar-Cederlof, I.; von Bahr, C.; Borga, O. and Sjoqvist, F.: Protein binding of diphenylhydantoin and desmethylirnipramine in plasma from patients with poor renal function. New England Journal of Medicine 825 : 264-267 (1971). Remmer, H.; Hirschmann, 1. and Greiner, I.: Die Bedeuting von Kumulation und Elimination fiir die Dosierung von Phenytoin (Diphenylhydantoin). Dtsch. Med. Wochenschr. 94 : 1265 - 1272 (1969) . Reynolds, E.H.: Longitudinal studies of serum antiepileptic drug levels preliminary observations : interaction of phenytoin and primidone; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.80-86 (Springer-Vedag, Berlin 1975). Reynolds, E.H.; Fenton, G.; Fenwick, P.; Johnson, A. L and Laundy, M. : Interaction of phenytoin and primidone. British Medical Journal 2: 594-595 (1975). Richens, A. and Dunlop, A.: Serum-phenytoin levels in management of epilepsy . Lancet 2: 247-248 (1975a). Richens, A. and Dunlop, A.: Phenytoin dosage nomogram. Lancet 2: 1305-1306 (1975b). Rieder, 1.: Binding of c!onazepam (active principle of Rivotril Roche) to the total proteins of mixed human plasma, as measured by equilibrium dialysis. Report No. 53'868, F. Hoffman-La Roche & Co. Ltd., Basel, Switzerland (1973). Rieder, J. and Wendt, G.: Pharmacokinetics and metabolism of the hypnotic nitrazepam; in Garattini, Mussini and Randall (Eds) The Benzodiazepines, p.99-128 (Raven Press, New York 1973). Schneider, H.: Carbamazepine: An attempt to correlate serum levels with anti-epileptic and side effects; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs p.15l - 158 (Springer-Verlag, Berlin 1975). Schobben, F.; Klein, E. van der and Gagreels, F.J.M.: Pharmacokinetics of di-n-propylacetate in epileptic

Clinical Pharmacokinetics of Anticonvulsants

patients. European Journal of Clinical Pharmacology 8: 97-105 (1975). Shannon, R.W.; Fraser, G.P.; Aitken, R.G. and Harper, J .R. : Diazepam in preeclamptic toxaemia with special reference to its effect on the newborn infant. British Journal of Clinical Practice 26 : 271 - 275 (1972). Sherwin, A.L and Robb, J.P.: Ethosuximide. Relation of plasma levels to clinical co ntrol; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.443-448 (Raven Press, New York, North-Holland, Arnst. 1972). Shoeman, D.W. ; Benjamin, D.M. and Azarnoff, D.L. : The alteration of plasma proteins in uremia as reflected in the ability to bind diphenylhydantoin. Annals New York Academy of Sciences 226 : 127- 129 (1973). Simonsen, N. ; Olsen, P.Z.; Kiihl, V.; Lund, M. and Wendelboe, I. : A double-blind study of carbamazepine and diphenylhydantoin in temporal lobe epilepsy. Acta Neurologica Scandinavica Suppl. 60 : 39 - 42 (1975). Sjo, 0.; Hvidberg, E.F.; Larsen, N.-E.; Lund, M. and Naestoft, J.: Dose-dependent kinetics of ethotoin in man. Clinical and Experimental Pharmacology and Physiology 2: 185 - 192 (1975a). Sjo, 0 .; Hvidberg, E.F.; Naestoff, J. and Lund, M. : Pharmacokinetics and side-effects of c10nazepam and its 7-amino-metabolite in man. European Journal of Clinical Pharmacology 8: 249- 254 (1975b). Sjoholm, I.; Kober, A.; Odar-CederlOf, I. and Borga, 0.: Protein binding of drugs in uremic and normal serum. The role of endogenous binding inhibitors. Biochemical Pharmacology (In press, 1976). Solomon, H.M. and Schrogie, J.J.: The effect of p. henyranudol on the metabolism of diphenylhydantoin. Clin. Pharmacol. Ther. 8: 554-556 (1967). Sonnen, A.E.H.; Blom, G.F. and Meijer, 1.W.A. : Enteric coated dipropylacetate (Depakine ®) ; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.229 - 234 (Springer-Verlag, Berlin 1975). Stewart, M.J.; Ballinger, B.R. ; Devlin, E.J.; Miller, A.Y. and Ramsay, A.C.: Bioavailability of phenytoin. A comparison of two preparations. European Journal of Clinical Pharmacology 9: 209-212 (1975). Strandjord, R.E. and Johannessen, S.l. : One daily dose of diphenylhydantoin for patients with epilepsy. Epilepsia 15: 317 - 327 (1974). Strandjord, R.E. and Johannessen, S.I. : A preliminary study of serum carbamazepine levels in healthy su~ jects and in patients with epilepsy; in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, p.1 81-188 (Springer-Verlag, Berlin 1975). Svendsen, T.L; Kristensen, M.B.; Hansen, J.M. and Skovsted, L: The influence of disulfiram on the half

188

life and metabolic clearance rate of diphenylhydantoin and tolbutamide in man. European Journal of ainical Pharmacology 9: 739 - 741 (1976) . Svensmark, O. and Buchthal, F.: Accumulation of phenobarbital in man. Epilepsia 4: 199-206 (I 963). Svensmark, O. and Buchthal, F.: Diphenylhydantoin and phenobarbital (serum levels in children). American Journal of Diseases in Children 108: 82 - 87 (1964). Troupin, A.S. and Friel, P.: Anticonvulsant level in saliva, serum and cerebrospinal fluid. Epilepsia 16: 223 - 227 (1975). Troupin, A.S.; Green, 1.R. and Halpern, LM.: Carbamazepine (Tegretol ®) as an anticonvulsant , Acta Neurologica Scandinavica Suppl. 60 : 13-26 (1975) . Troupin, A.S.; Green, J.R. and Levy, R.H. : Carbamazepine as an anticonvulsant. A pilot study. Neurology 24: 863-869 (1974). Vajda, F.; Williams, F.M.; Davidson, S.; Falconer, M.A. and Breckenridge, A.: Human brain, cerebrospinal fluid, and plasma concentrations of diphenylhydantoin and phenobarbital. ainical Pharmacology and Therapeutics 15: 597 - 603 (1974). Waddell, W.J. and Butler, T.C.: The distribution and excretion of phenobarbital. Journal of ainical Investigation 36: 1217-1226 (1957). Wilder, B.l . and Ramsay, R.E. : Oral and intramuscular phenytoin. ainical Pharmacology and Therapeutics 19: 360- 364(1976). Wllder, B.l.; Serrano, E.E. ; Ramsey, E. and Buchanan, R.A. : A method for shifting from oral to intramuscular diphenylhydantoin administration. Clinical Pharmacology and Therapeutics 16 : 507 - 513 (1974). Wilder, B.1. ; Streiff, R.R. and Hammer, R.H.: Diphenylhydantoin. Absorption, distribution and excretion. Clinical studies ; in Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs, p.137- 148 (Raven Press, New York, North-Holland, Amsterdam 1972). Woodbury, D.M. ; Penry, 1.K. and Schmidt, R.P. : (Eds) Antiepileptic Drugs (Raven Press, New York ; NorthHolland, Amsterdam 1972). Yonekawa, W.; Kupferberg, H.l . and Cantor, F. : A gas chromatographic method for the determination of ethotoin (3-ethyl-5-phenylhydantoin) in human plasma; in Schneider, lanz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Pharmacology of AntiEpileptic Drugs, p.llS-121 (Springer-Verlag, Berlin 1975). Author's address: Prof. Eigill F. Hvidberg, Clinical Pharmacology Research Unit, Rigshospitalet (University Hospital) 9 Blegdamsvej, DK-2100 CopenJuzgen Q (Denmark).