Bertilsson 1978
Bertilsson 1978
Bertilsson 1978
Leif Bertilsson
Department of Clinical Pharmacology at the Karolinska Institute, Huddinge Hospital, Hud-
dinge .
Summary Carbamazepine seems to be as effective as phenytoin in the treatment of grand mal and psy-
chomotor epilepsy. It is the drug of first choice in trigeminal neuralgia.
After single oral doses of carbamazepine, the absorption isfairly complete and the elimina-
tion half-life is about 35 hours (range IB to 65 hours)' During multiple dosing, the half-life is
decreased to 10-20 hours, probably due to autoinduction of the oxidative metabolism of the
drug. Phenytoin and barbiturates also induce the metabolism of carbamazepine. After single
doses of carbamazepine, elimination follows dose-dependent first order kinetics.
Carbamazepine is metabolised by oxidation before excretion in the urine. In experimental
animals, the metabolite carbamazepine-IO, II-epoxide has anticonvulsant activity comparable
with thaLof the parent drug. The plasma concentration of the metabolite during 'Iong:tetfif
treatment of epileptic patients varies between 5 and Bl % of that of the parent drug. The
plasma protein binding of the metabolite is about 50 % compared with about 75 % for the
parent drug. Less than 50 % of a given carbamazepine dose has been identified as metabolites
in the urine. The quantitatively most important metabolite is the trans-I O,ll-dihydro-I 0,11-
'dial.
The kinetics of carbamazepine have been explored to some extent in pregnant women,
newborns and children. Plasma levels of carbamazepine seem to decrease during pregnancy,
possibly as a result of increased metabolism. The drug readily crosses the placellta and the
levels measured in newborns are comparable with maternal plasma concentrations. In
newborns exposed to the drug during fetal life, the plasma half-lives were relatively short (B.2
to 2B.I hours) indicating an induction of carbamazepine metabolism during gestation. The
pharmacokinetiCS of carbamazepine in children aged 0.3 to 15 years are comparable with that
in adults.
A single daily dose of carbamazepine is insufficient; 2 doses per day are appropriate in most
cases, but some patients may benefit from more frequent dosing to avoid side-effects.
Compared with phenytoin, for example, very few controlled studies have been performed to
establish the plasma level range of carbamazepine associated with the best therapeutic out-
come. However, the best anticonvulsant effect of carbamazepine seems to be obtained at plasma
levels of about 5 to IOJlg/ ml (20 to 40Jlmol/ L). Side-effects are most frequent at higher levels
but may also be seen at lower levels.
Clinical Pharmacokinetics of Carbamazepine 129
/
16
14
Liquid
chromatography
.
12
10
.1·
8
.... .
•
//
I,,{/'" •
/
~f' 0
c::
o
';::;
~c::
6
.-...
.-. ".
.,)
~ .
;JrtoS"
Q)
~ 4
o
"'"
E
'"'"
0:: 2 ~/
.~
o /-
/ Mass fragmentography
o 2 4 6 8 10 12 14 16
Fig. 1. Comparison of a liquid chromatographic (Eichelbaum and Bertilsson, 1975) and a mass fragmentographic method
(Palmer et aI., 1973) for the determination of carbamazepine in plasma.
There was a significant linear regression correlation of the concentrations (fig/mil obtained by the 2 methods when plasma
from patients treated long-term with the drug was analysed (n = 108; r = 0.99; p < 0.001). The regression line y = 0.98x +
0.28 is indicated by a solid line; the dashed line is y = x (from Eichelbaum and Bertilsson, 1975).
day following ingestion of about 20g (Gruska et aI., (in gelatin capsule) is more than 70 % (Faigle and
197 n. Feldmann, 1975). The areas under the plasma con-
The absolute bioavailability of carbamazepine in centration-time curve (AUC) extrapolat~ to infinite
man has not been determined, because a preparation time are similar when the drug is given as a tablet or
for intravenous injection is not available. However, as a solution in aqueous ethanol (Rawlins et aI., 1975)
based on the recovery of radio-labelled carbamazepine or in propylene glycol (Levy et aI., 1975b). The AUC
in urine and faeces, the oral bioavailability of the drug after a single dose of carbamazepine is proportional to
Clinical Pharmacokinetics of Carbamazepine 131
Rawlins et al. (1975) 8 200 Ethanol soln 35.9 ± 8.3 1.07 ± 0.22
(23.9 - 46.4) (0.79 - 1.40)
the dose given within the range SO to 600mg solubilisation of the drug by the bile secreted follow-
(Rawlins et aI., 1975; Levy et aI., 1975b; Gerardin ing the meal.
and Hirtz, 1976). The bioavailability of car-
bamazepine tablets was increased when taken 2. J.2 Distribution
together with meals (Levy et aI., 1975b). The in- Carbamazepine is a neutral and fairly lipophilic
vestigators suggest that this is due to an increased compound that easily passes through membranes in
Clinical Pharmacokinetics of Carbamazepine 132
Johannessen and Strandjord (1972) CSF-plasma ratio Epileptic patients ( 19) 21.0 ± 5.0
Rawlins et al. (1975) Equilibrium dialysis Normal subjects (8) 27.9 ± 0.4 (26.0-29.8)
at 37°C
McAuliffe at al. (1977) Equilibrium dialysis Epileptic patients ( 12) 25.9 ± 3.4
at 24°C
Saliva-plasma ratio Epileptic patients ( 12) 26.0 ± 2.4
Pynnonen et al. (1977) CSF-plasma ratio Epileptic children (9) 33.4 ± 15.7
the body. In animals, the drug penetrates into all drug and the real figures may therefore be slightly
tissues and organs within I minute after an in- lower.
travenous injection (Faigle et ai., 1976). The distribu-
tion pattern does not suggest that the drug has any 2./.3 Plasma Protein Binding
preferential affinity for particular organs (Morselli et The plasma protein binding of carbamazepine has
ai., 1971; Faigle et ai., 1976). In man, carbamazepine been investigated using in vitro techniques (equi-
concentrations have been examined in such tissues librium dialysis and ultrafIltration) and in vivo by
and fluids as brain (Morselli, 1975; Morselli et ai., estimating the concentration of the drug
1977), cerebrospinal fluid (Johannessen et ai., 1976), simultaneously in cerebrospinal fluid (or saliva) and
plasma of newborn infants after a transplacental ex- in plasma (table 11). There is a rather good agreement
posure of the fetuses (Rane et ai., 1975) and mother's between the different studies and the protein binding
breast milk (pynnonen and Sillanpaa, 1975). of carbamazepine in most cases is 70 to 80 % . The as-
After single oral doses of carbamazepine to volun- sociation constant for albumin is low and other
teers or patients (table I), the apparent volume of dis- plasma proteins may be implicated in the binding of
tribution has been found to vary between 0.79 and carbamazepine (DiSalle et ai., 1974).
1.86L/kg body weight. These figures have been The free (non-protein bound) fraction of the drug
calculated assuming complete bioavailability of the in plasma is in equilibrium with the drug at the recep-
Clinical Pharmacokinetics of Carbamazepine 133
tors where it exerts its effect. Thus, it is this fraction generally are toxic compounds (cJ Frigerio et aI.,
of the drug in plasma that should be correlated with, 1976). Carbamazepine-I 0, I I-epoxide (II) is,
for example, clinical effects. On the other hand, if however, a comparatively stable epoxide and seems to
there is little variation in the degree of protein bind- be no more toxic than carbamazepine itself (Glatt et
ing, the total plasma level reflects the free fraction. In aI., 1975; cJ Frigerio et aI., 1976).
the material of Hooper et al. (J 975) [table Ill, the un- Metabolites II (Frigerio et al., 1972) and III
bound fraction of carbamazepine showed a 4-fold (Goenechea and Hecke-Seibicke, 1972; Baker et aI.,
variation between individuals. This variation was less 1973a; Faigle and Feldmann, 1975) have been isol-
pronounced in the other studies. ated from human urine and their structures deter-
As the free drug concentration. is much more mined, although both compounds rearrange in the
difficult to measure than the total level, the cp.oice bet- gas chromatograph to 9-acridinecarboxyaldehyde
ween these two in monitoring plasma levels depends (Baker et aI., 1973b; Frigerio and Morselli, 1975).
on available laboratory facilities. Recently, it has been The configuration of the diol (III) formed in man is
suggested that the concentration of carbamazepine in trans and it is optically active, which shows that III is
saliva reflects the free fraction in plasma (Chambers formed by an enzymatic ring opening of the epoxide
et aI., 1977; McAuliffe et aI., 1977) and if future on [Faigle and Feldmann, 1975]. This transformation
studies show a wide variation in protein binding, this is catalysed by an epoxide hydrase (Oesch, 1973).
noninvasive technique may be valuable to use. After single doses of carbamazepine to volunteers, ap-
Hooper et al. (J 975) found that patients with proximately I % of the dose is found in urine as the
hepatic disease had a significantly lower protein bind- epoxide (II) and about 20 % as the diol (III), about a
ing than controls, but the difference was very small third of which is conjugated, presumably as the
and is of no practical importance. Renal disease had glucuronide (Faigle and Feldmann, 1975). The form-
no influence on the binding. Drugs such as phenytoin ation of iminostilbene (metabolite V; Csetenyi et aI.,
and phenobarbitone that are frequently used together 1973; Morselli et aI., 1975), carbamazepine hydrox-
with carbamazepine have no influence on the protein ylated in the aromatic nucleus (metabolite IV; Faigle
binding of carbamazepine (Hooper et aI., 1975; Mor- et aI., 1976) and metabolite VI (Faigle et al., 1976)
selli et aI., 1975; Rawlins et aI., 1975). seem to be quantitatively less important pathways in
the metabolism of carbamazepine. Most of the urin-
2.1.4 Metabolism ary metabolites of carbamazepine have thus not yet
After the administration of a single oral dose of been identified (Faigle and Feldmann, 1975; Faigle et
'4C-carbamazepine to 2 volunteers, 72 % of the ad- .aI., 1976). It can not be excluded that the 20 % found
ministered radioactivity was excreted in the urine and as II and III is an underestimate of the epoxide path-
the rest in faeces over a period of 9 days (Faigle and way in the metabolism of carbamazepine. The epox-
Feldmann, 1975). The slow appearance of radioac- ide may be further metabolised along not yet iden-
tivity in faeces suggests that it had been absorbed in tified pathways (e.g. conjugation with glutathion),
the gastrointestinal tract and then excreted via the although further metabolism seems to be slow as
bile. judged from the clearance of the epoxide in the isol-
Only minor quantities of unchanged car- ated perfused rat liver (Rane et aI., I 977).
bamazepine (about 2 % of the dose) is found in
human urine (Faigle and Feldmann, 1975; Morselli et 2.1.5 Elimination Kinetics
aI., 1975). Most of the drug is metabolised and 2 ox- In contrast to the short half-lives (J to 2h) of car-
idative pathways have been recognised in man. (fig. bamazepine reported in the rat (Farghali-Hassan et
2). Major interest has been focussed on the epoxide- aI., 1976)and in the monkey (Levy et ai., 1975a) after
diol pathway (metabolites II and III), as epoxides single doses, we found half-lives ranging from 24 to
Clinical Pharmacokinetics of Carbamazepine 134
-
/
~
tL~.,~)
~ 'j
c=o
I
NH,
~ - II~
A .. h -jJ-
~ N·
I
c=o
I
NH,
~
IV
OH
I
I
I
1
Conjugates
H CH,OH
C(X)
"
1.<:: ;,...1
N
I
:;...-
c=o
I
... ...
NH, VI
- - Conjugates
Conjugates
46 hours in human volunteers (Palmer et al., 1973; 1975). Apparently, the pharmacokinetics of car-
Rawlins et al., I 975).This range of half-lives has also bamazepine change during multiple dosing.
been found in other studies in pteviousl~ untreated To. investigate this phenomenon, we administer~_ .
volunteers or patients (table I). a single oral dose of carbamazepine to 4 patients
The elimination of carbamazepine follows dose-in- (Eichelbaum et al., 1975) and found plasma half-lives
dependent kinetics (Levy et al., 1975b; Rawlins et al., of 35.6 ± 15.3 hours (table III). The patients were
1975). Mean plasma clearances of carbamazepine in 2 subsequently treated with 200mg of carbamazepine 3
studies (Eichelbaum et al., 1975; Rawlins et al., times a day for 15 to 21 days. The steady state plasma
197 5) were 30 and 25 mll min, respectively. This levels that were achieved were lower than would be
drug has a much lower clearance than drugs such as predicted from the single-dose experiments. The
lignocaine (lidocaine), nortriptyline and propranolol levels seemed to decrease during the treatment period.
(cf Rawlins et al., 1975) and a 'rlIst-pass' hepatic When the treatment was stopped the post-steady state
metabolism is highly unlikely. plasma half-lives (20.9 ± 5.0 hours) were shorter
The kinetics of the drug after single doses should than those obtained after the single dose (ftg. 3 and ta-
be able to be used for prediction of steady state plasma ble III). The highest levels of the drug that were ob-
levels during multiple dosing (Wagner et al., 1965; tained during the multiple dosing (8.4 ± 1.6J.lg/mO
Alexanderson, 1972), but the predicted levels thus ob- were lower than those predicted from the single dose
tained are 2 to 3 times higher than the concentrations experiments (I 7.2 ± 7.2 J.lg I mO. The results obtained
reported in patients undergoing long-term treatment in this study (Eichelbaum et al., 1975) suggest that
with carbamazepine at that dose (Rawlins et aI., carbamazepine induces its own metabolism (auto in-
Clinical Pharmacokinetics of Carbamazepine 135
duction). Similar conclusions from other studies bamazepine and the epoxide metabolite. The mean
(table III) have been drawn and Pitlock et al. (1976) fluctuation of carbamazepine in plasma of patients
have developed a pharmacokinetic model to describe receiving the drug 2 and 3 times daily was similar (57
the self-induced decrease in steady state concentra- ± 20 and 56 ± 29%, respectively), while 4 daily
tions of carbamazepine. Studies in epileptic patients doses reduced the fluctuation to 36 ± 17 %. These
(Morselli et aI., 1975) show that the post steady state authors conclude that a single daily dose of car-
plasma half-lives are similar to the half-lives after bamazepine is insufficient. 2 doses per day· are ap-
multiple doses reported in table III. propriate in most cases, but some patients may
benefit from more frequent dosing to avoid side-
2.2 Plasma Levels of Carbamazepine During effects.
Long-Term Treatment
20
10
5 ~~------.------~
.--.:\
\
05
02
•
0:1 \
0.05
0 5 10 15 20 25 o 50 100 150
Days Hours
Fig. 3. Plasma concentrations (pg/ml; ordinate) of carbamazepine ( • ) and the epoxide metabolite ( * ) determined at various
times (abscissa) in patient RA.
The patient first received a single oral 200mg dose' of carbamazepine (broken lines in the right hand graph). After 1 week the
patient received 200mg of the drug 3 times a day for 21 days (left hand graph). Arrow indicates when the last dose was given.
Unbroken lines in the right hand graph show the plasma concentration of the parent drug and the epoxide after the last dose in
the multiple dose regimen.
The theoretical steady state plasma concentration of carbamazepine during multiple doses shown in the figure (...........l
(23.6pg/mll was predicted from pharmacokinetic data obtained in the single dose experiments (from Eichelbaum et aI., 1975).
Clinical Pharmacokinetics of Carbamazepine 136
Table III. Pharmacokinetic properties of carbamazepine after single and multiple oral doses of Tegretol tablets [Mean ± SD
(range)]
Eichelbaum et al. (1975) 4 35.6 ± 15.3 17.2 ± 7.2 200 x 3 15 - 21 20.9 ± 5.0 8.4 ± 1.6
(18.5 - 54.7) (7.0 - 23.6) (16.4 - 26.6) (6.3 - 9.9)
Faigle et al. (1976) 6 36.9 ± 3.2 200 x 1 14 27.9 ± 1.9 3.4 ± 0.2
(28.9 - 40.8) (27.7 - 36.5) (2.9 - 4.3)
Pitlick et al. (1976) 6 33.9 ± 3.5 Twice that 400 x 1 22 19.8 ± 4.0
obtained
bamazepine induces its own metabolism among reported by Hooper et al. (1974) are the only ones in-
different subjects, suggest that the plasma concentra- dicating dose-dependent kinetics of carbamazepine
tion achieved cannot be predicted from the dose and remain to be conftrmed.
prescribed. In fact, no significant correlation (n = 93;
r = 0.19) was observed between the prescribed dose 2.2.3 Combination With Other Antiepileptic Drugs
and the plasma concentration of carbamazepine in 25 The plasma levels of carbamazepine were higher
epileptic patients treated for at least 6 months with when patients were given carbamazepine alone com-
the drug (Eichelbaum et at., I 976a.,b). Such a poor pared with combined treatment with either phellytoin.
correlation has also been shown by other investiga- or phenobarbitone (Christiansen and Dam, 1973;
tors (Meinardi, 1972; Christiansen and Dam, 1973; Johannessen and Strandjord, 1975; Schneider,
Hooper et aI., 1974; Johannessen and Strandjord, I 975b). Treatment with all 3 drugs produced the
1975; Schneider, 1975b; Morselli et ai., 1976). lowest plasma levels of carbamazepine. The most
Meinardi (1972) showed in 2 patients that doub- probable explanation for this interaction is an induc-
ling the dose of carbamazepine doubled the plasma tion of the liver microsomal drug metabolising
level of the drug. Hooper et aI. (1974) measured car- enzymes, although decreased absorption (or drug in-
bamazepine plasma levels at 2 or more dose levels in take) cannot be excluded.
10 patients, with measurements performed long
enough after the dosage changes for the new steady 2.2.4 Pregnancy
state to be established. The increases of the dose pro- Christiansen and Dam (I977) measured car-
duced a disproportionally high increase of the plasma bamazepine and the epoxide metabolite in plasma of
level, suggesting that the relationship between plasma epileptic patients before, during and after pregnancy.
carbamazepine and dose in the individual patient is The level of carbamazepine decreased, While that of
curvilinear. Such a relationship indicating dose-de- the epoxide increased at the peripartal time. These
pendent kinetics has previously been shown for results indicate an increased clearance of car-
phenytoin (Richens and Dunlop, 1975). The' results bamazepine during pregnancy, a finding also ob-
Clinical Pharmacokinetics of Carbamazepine 137
served with phenytoin and phenobarbitone (Eadie et those measured in adults after multiple doses of car-
aI., 1977). In accordance with this, we found low bamazepine, but shorter than those obtained after
concentrations of carbamazepine in plasma of epilep- single doses (table III). This indicates that these
tic women at the time of delivery (Rane et a\., 1975) newborn babies had a well-developed metabolic
compared with non-pregnant epileptics (Eichelbaum capacity, possibly through transplacental autoinduc-
et a\., I 976a). tion (Rane et a\., 1975). Recent studies (Piafsky and
Rane, unpublished observations) have shown that
human fetal livers are able to metabolise car-
2.3 Pharmacokinetic Drug Interactions bamazepine to the epoxide in vitro.
Pynnonen and Sillanpaa (I975) showed in I
Phenytoin and phenobarbitone seem to induce the breast-feeding mother that carbamazepine is present
metabolism of carbamazepine, thereby decreasing the in breast milk at a level of about 50 % of that in
plasma level of this drug (Christiansen and Dam, plasma. After 30 days of unaltered medication and
1973~ Johannessen and Strandjord, 1975; Schneider, nurs'ing, the neonate's plasma concentration was
1975b). Other drugs such as propoxyphene (Dam and more than half of the mother's plasma concentration.
Christiansen, 1977) and triacetyloleandomycin
(Dravet et aI., 1977) increase the plasma levels of car-
bamazepine, probably by inhibiting its hepatic 3.2 Children
metabolism. As carbamazepine induces its own
metabolism, it is not surprising that it also can induce The steady state plasma concentration of car-
the metabolism of other drugs; e.g. phenytoin and bamazepine was assessed in 43 epileptic children (2 to
warfarin (Hansen et aI., 1971), the tetracycline doxy- 15 years of age) on carbamazepine treatment (Rane et
cycline (Penttila et aI., 1974; Neuvonen et aI., 1975) a\., 1976). The levels were in the same range as had
and probably oraI hormonal steroid contraceptives been found in adult patients on corresponding doses.
(Hempel and Klinger, 1976). A correlation (r = 0.65) was noted between dose and
plasma level in the group of children on single-drug
treatment, but there was no correlation in the group
3. Pharmacokinetics oJCarbamazepine in of children on a combined drug therapy regimen.
Newborns and Children Plasma half-lives of carbamazepine estimated after'
discontinuation of therapy in 2 children were 13.7
3.1 Newborn Infants and 18.9h. This indicates an autoinduction of the
metabolism of carbamazepine in these children.
A single dose of carbamazepine to pregnant rats Children on combined anticonvulsant treatment had
equilibrates across the' placenta within 30 minutes lower carbamazepine levels and, expressed as a per-
after drug administration (Farghali-Hassan et aI., centage of the parent drug, the metabolite concentra-
1975, 1976). We have studied the disposition of tion was higher than in children treated only with
transplacentally transferred carbamazepine in 5 carbamazepine.
newborn infants (Rane et a\., 1975). The mothers had Morselli (I975) found no relationship between
received carbamazepine and phenytoin for sound daily dose and plasma level of carbamazepine in
therapeutic indications throughout gestation. The epileptic children. This lack of correlation was most
plasma concentrations of carbamazepine in the evident in cases where other anticonvulsants were
newborns and others were comparable at birth. The also administered. Pynnonen et al. (I 977) also found
plasma elimination half-lives in these 5 newborns a weak correlation (r = 0.56) between dose and
(8.2, 10.5, 10.8, 27.7 and 28.lh) were similar to plasma level.
Clinical Pharmacokinetics of Carbamazepine 138
4. Carbamazepine-/O,ll-Epoxide After the 1977) and may be an index of the degree of metabolic
Administration ofCarbamazepine- induction.
Pharmacological and Toxicological Aspects In a very recent study by Johannessen et al.
(1977), the plasma levels of carbamazepine and the
As discussed above (section 2.1.4), carbamazepine epoxide were determined over I day following dosing
is metabolised to the 10, II-epoxide (II in fig. 2). This 2 to 4 times daily. The level of the epoxide tended to
metabolite is fairly stable and it has been identified in increase with increasing number of daily doses.
human urine (Frigerio et aI., 1972), plasma The epoxide has not been given to man and its
(Eichelbaum and Bertilsson, 1975), and brain (Mor- kinetics have therefore not been fully evaluated. After
selli et al., 1977). The epoxide metabolite can be multiple dosing of carbamazepine, the post-steady
measured by either gas chromatography (Morselli et state half-lives of the epoxide metabolite are slightly
al., 1973) or high-pressure liquid chromatography shorter than those of the parent compound in each in-
(Eichelbaum and Bertilsson, 1975). In the former dividual (Eichelbaum et aI., 1975; Morselli et aI.,
method, a rearrangement product (9-acridine- I 976). The reported half-lives of the epoxide (range 6
carboxylaldehyde; Baker et aI., 1973b) formed in the to 23h) may be longer than the actual ones since
gas chromatograph is quantitated. metabolite formation continuously influences the
Chemically synthesised carbamazepine-I 0,11- elimination curve.
epoxide has an anticonvulsant activity similar to the The plasma protein binding of the epoxide is lower
parent drug when administered to animals (Frigerio than that of carbamazepine, determined by cerebro-
and Morselli, 1975; Gagneux, 1976). The relation- spinal fluid-plasma measurements (Eichelbaum et aI.,
ship between the anticonvulsant effect and the blood 1976a; Johannessen et aI., 1977; Pynnonen et aI.,
concentration of carbamazepine and for the epoxide 1977) and ultrafiltration (Morselli et al., 1975;
following the administration of either the parent drug Johannessen et aI., 1977). The CSF levels of the epox-
or the metabolite has. been investigated by Faigle et al. ide are about 50 % of the plasma levels, while the cor-
(J 977). Their observations suggest that car- responding figure for carbamazepine is only about
bamazepine possesses an anticonvulsant effect of its 25 % (table ro. The concentration of non-protein
own which is added to by the metabolite. bound epoxide and. carbamazepine are sS'llletirnes
The levels of the epoxide in plasma from epileptic similar during long-term treatment and it seems im-
patients treated long-term with carbamazepine are portant to measure both compounds when biological
poorly correlated with the given dose, both in adults (therapeutic, toxic, etc) effects are correlated with
(Eichelbaum et aI., I 976a,b) and in children (Rane et plasma levels.
aI., 1976). The plasma concentration of car-
bamazepine and the metabolite are better correlated,
but there is a pronounced variation in the ratio of
metabolite to parent drug (range 15 to 55 % in adults
and 5 to 8 I % in children). Similar results have been 5. Plasma Levels oiCarbamazepine in
obtained by other investigators (Christiansen. and Relation to Therapeutic Outcome and Side-
Dam, 1975; Schneider, 1975b; Morsellietal., 1975). Effects
The ratio between the epoxide and parent drug in-
creases when patients are treated with other anti- During the last few years the clinical phar-
epileptic drugs such as phenytoin and phenobarbitone macokinetics of carbamazepine have been thoroughly
(Christiansen and Dam, 1975; Schneider, 1975b; investigated (see above), but very few controlled clini-
Rane et aI., 1976). This ratio was similar in children cal trials have been performed relating plasma levels
and in adults (Rane et aI., 1976; Pynnonen et al., to therapeutic and toxic effects. When evaluating the
Clinical Pharmacokinetics of Carbamazepine 139
significance of such studies several factors have to be best control of seizures was obtained when patients
considered, for example: were treated with all 3 drugs. An optimum dose of
I) Were the studies performed in patients who did carbamazepine was reported to be 1200mg daily but
not respond to other anti epileptic drugs or were no clear relationship between plasma level and effect
they not previously treated? Were the patients could be demonstrated.
treated concomitantly with other drugs; e.g. Monaco et al. (I976) in 20 epileptic patients
phenytoin? Were the plasma levels of these unresponsive to treatment, also measured plasma
other drugs within the therapeutic range? levels of carbamazepine, phenytoin and phenobar-
2) Were the plasma levels measured on several bitone. A decrease in seizure frequencies was attained
occasions to establish that the patients were in at plasma levels of 4 to 10pg of carbamazepinel ml.
steady state? Variations in the plasma level Carbamazepine-I 0, I I-epoxide was also monitored
may suggest poor compliance with respect to but no relationship to effect could be shown.
the patient's drug-intake. Dam et al. (t 975) found carbamazepine to be
3) Was the epoxide metabolite (probably active) effective in grand mal epilepsy at levels higher than
also measured in plasma? 4).1g/ml plasma. Carbamazepine given alone was
4) Was the analytical method specific and has this rather ineffective in the treatment of psychomotor
been established by comparison with other seizures. The concentration of the epoxide metabolite
techniques? in plasma was increased when the patients were
5) Was the plasma sample drawn just before the treated with carbamazepine in combination with
morning dose (and not after)? phenytoin and/ or phenobarbitone. The authors sug-
6) Was the therapeutic effect investigated at gest that this may explain the better clinical effect of
different plasma levels (doses) within the same this combined treatment as the epoxide has an inde-
patient? pendent anti epileptic effect.
7) Was the study performed blind? Was the clini- Troupin et al. (1975) indicated a therapeutic level
cal effect assessed with a proper technique? of 7 to 16).1g carbamazepine/ ml plasma in the treat-
8) Was the type of epilepsy clearly stated? ment of grand mal epilepsy.
A short review on some studies aimed to correlate Simonsen et al. (1976) in a double-blind crossover
plasma levels with biological effects follows. study found that carbamazepine and phenytoin have a
Schneider (197 Sa) showed that the serum level of similar effect in preventing psychomotor seizures.
carbamazepine was positively correlated with the Some patients responded better to carbamazepine
degree of seizure control. In a group of institu- than to phenytoin while in other patients the opposite
tionalised patients, complete control of psychomotor was true. The therapeutic range of carbamazepine
seizures was obtained at a mean level of 4.6 ± 1.3pg was said to be 6 to I O).lg/ ml.
carbamazepinel ml serum. The corresponding figure In another study, patients with partial epilepsy
in recently treated patients was 6.5 ± 3.0pgl ml. The with complex symptomatology, who were already
same author found in 10 patients that the minimal being treated with phenytoin in optimum doses
mean level to cause side-effects was 11.6 ± (plasma levels 14 to 20).lgl mO were also given car-
4.3J..ig/ml, with a threshold serum level for more bamazepine (Eichelbaum et al., 1976a). Both car-
serious and pronounced side-effects (distinct drowsi- bamazepine and the epoxide metabolite were meas-
ness and slowing, ataxia and diplopia) at about ured in plasma. At a level of 5).1g carbamazepine/ml
9).1g/ml. plasma there was no further reduction in the frequen-
Cereghino and colleagues (1973, 1974, 1975) cy of partial or generalised epileptic seizures. In 5
have compared the effects of carbamazepine, pheny- patients, the dose of carbamazepine was increased to
toin and phenobarbitone in double-blind studies. The produce a plasma concentration of 7 to 8).1g/ml.
Clinical Pharmacokinetics of Carbamazepine 140
There was still no improvement, but side-effect" were effect" are most frequent at higher levels, but may
seen in 3 patients. also be seen at lower levels.
Most of the studies discussed above have design It seems appropriate to recommend plasma level
and methodological disadvantages. However, despite monitoring of carbamazepine in the treatment of
these drawbacks the studies indicate that the epilepsy. Further controlled studies in selected groups
therapeutic plasma level of carbamazepine is about 5 of patients are also needed to establish optimum
to 10).lg/ml (20 to 40).lmol/L) in the treatment of plasma levels of carbamazepine and of the active
grand mal or psychomotor epilepsy. The epoxide epoxide metabolite.
metabolite may also contribute to the therapeutic
effect, but it is not known to what degree.
Side-effects are usually most pronounced during Acknowledgements
the initiation of therapy and therefore when the dose
is gradually increased (e.g. Gamstorp, 1975). This The studies performed in our laboratories were supported by
grants from the Swedish Medical Research Council (04X-03902),
may be explained by our findings that the highest
Stiftelsen Margaretahemmet and the Karolinska Institute.
plasma concentrations of both carbamazepine and the
epoxide are observed after 3 to 4 days and subse-
quently seem to decrease because of an autoinduction
effect (Eichelbaum et aI., 1975; see also fig. 3).
References
Christiansen, J. and Dam, M.: Influence of phenobarbital and Ekbom, K.: Carbamazepine in the treatment of tabetic lightning
diphenylhydantoin on plasma carbamazepine levels in patients pains. Archives of Neurology 26: 374-378 (I 972).
with epilepsy. Acta Neurologia Scandinavica 49: 543-546 Faigle, J.W.; Brechbuhler, S.; Feldmann, K.F. and Richter, WJ.:
(I 973). The biotransformation of carbamazepine; in Birkmayer (Ed)
Christiansen, J. and Dam. M.: Drug interaction in epileptic pa- Epileptic Seizures - Behaviour - Pain, p.127-140 (Huber,
tients; in Schneider. Janz. Gardner-Thorpe. Meinardi and Berne 1976).
Sherwin (Eds) Clinical Pharmacology of Anti-Epileptic Drugs, Faigle, J.W. and Feldmann. K.F.: Pharmacokinetic data of car-
p.197-200 (Springer-Verlag, Berlin 1975). bamazepine and its major metabolites in man; in Schneider,
Christiansen, J. and Dam, M.: Plasma and salivary levels of car- Janz, Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical
bamazepine and carbamazepine-I 0.1 I-epoxide during preg- Pharmacology of Anti-Epileptic Drugs, p.159-165 (Springer-
nancy; in Gardner-Thorpe, Janz, Meinardi and Pippenger Verlag. Berlin 1975).
(Eds) Antiepileptic Drug Monitoring, p.128-137 (Pitman Faigle, J.W.; Feldmann. K.F. and Baltzer, V.: Anticonvulsant
Press, Avon, England 1977). effect of carbamazepine. An attempt to distinguish between the
Csetenyi, ].; Baker. K.M.; Frigerio, A. and Morselli, P.L.: Im- potency of the parent drug and its epoxide metabolite; in
inostilbene - a metabolite of carbamazepine isolated from rat Gardner-Thorpe, Janz, Meinardi and Pippenger (Eds) Anti-
urine. Journal of Pharmacy and Pharmacology 25: 340-341 epileptic Drug Monitoring, p.1 04-1 09 (Pitman Press, A von.
(I 97 3). England 1977).
Dam, M.' and Christiansen, J.: Interaction of propoxyphene with Farghali-Hassan, Assael, B.M.; Bossi, L.; Garattini, S.; Gerna, M.;
carbamazepine. Lancet 2: 509 (t 977). Gomeni, R. and Morselli. P.L.: Carbamazepine phar-
Dam, M.; Jensen, A. and Christiansen, J.: Plasma level and effect macokinetics in young. adult and pregnant rats. Relation to
of carbamazepine in grand mal and psychomotor epilepsy. pharmacological effects. Archives internationales de Phar-
Acta Neurologica Scandinavica, Suppl. 60: 33-38 (I 975). macodynamie et de therapie 220: 125-139 (1976).
DiSalle, E.; Pacifici, G.M. and Morselli, P.L.: Studies on the Farghali-Hassan, Assael. B.M.; Bossi, L. and Morselli, P.L.: Pla-
plasma protein binding of carbamazepine. Pharmacological cental transfer of carbamazepine in the rat. Journal of Phar-
Research Communications 6: 193-202 (I 974). macy and Pharmacology 27: 956-957 (1975).
Dravet, c.; Mesdjian, E.; Cenraud, B. and Roger, J.: Interaction Frigerio, A.; Baker, K.M. and Morselli, P.L.: Gas Chroma-
between carbamazepine and triacetyloleandomycin. Lancet I: tographic-mass spectrometric studies on carbamazepine; in
810-811 (1977). Costa and Holmstedt (Eds) Advances in Biochemical Psy-
Eadie, M J.: Plasma level monitoring of anticonvulsants. Clinical chopharmacology, vol. 7, p.125-134 (Raven Press, New York
Pharmacokinetics I: 52-66 (1976). 1973).
Eadie, MJ.; Lander, C.M. and Tyrer, J.H.: Plasma drug level Frigerio. A.; Cavo-Briones, M. and Belvedere, G.: Formation of
monitoring during pregnancy. Clinical Pharmacokinetics 2: stable epoxides in the metabolism of tricyclic drugs. Drug
427-436 (1977). Metabolism Reviews 5: 197-218 (J 976).
Eichelbaum. M. and Bertilsson, L.: Determination of car- Frigerio, A.; Fanelli, R.; Biandrate, P.; Passerini, G.; Morselli,
bamazepine and its epoxide metabolite in plasma by high-speed P.L. and Garattini, S.: Mass spectrometric characterisation of
liquid chromatography. Journal of Chromatography 103: carbamazepine-I 0, II-epoxide. a carbamazepine metabolite
135-140 (1975). isolated from human urine. Journal of Pharmaceutical
Eichelbaum, M.; Bertilsson, L.; Lund, L.: Palmer, L. and Sjoqvist, Sciences 61: 1144-1147 (I 972).
F.: Plasma levels of carbamazepine and carbamazepine-I 0.11- Frigerio, A. and Morselli, P.L.: Carbamazepine: Biotransforma-
epoxide during treatment of epilepsy. European Journal of tion; in Penry and Daly (Eds) Advances in Neurology, vol I I,
Clinical Pharmacology 9: 417-421 (I 976a). p.295-308 (Raven Press, New York 1975).
Eichelbaum, M.; Bertilsson, L.; Rane, A. and Sjoqvist, F.: Autoin- Gagneux. A.R.: The chemistry of carbamazepine; in Birkmayer
duction of carbamazepine metabolism in man; in Richens and (Ed) Epileptic Seizures - Behaviour - Pain, p.120-126
Woodford (Eds) Anticonvulsant Drugs and Enzyme Induc- (Huber, Berne 1976).
tion, p.14 7-158 (Associated Scientific Publishers, Amsterdam Gamstorp, I.: Treatment with carbamazepine: Children; in Penry
I 976b). and Daly (Eds) Advances in Neurology, vol II, p.237-246
Eichelbaum, M.; Ekbom, K.; Bertilsson, L.; Ringberger, V.A. and (Raven Press, New York 1975).
Rane, A.: Plasma kinetics of carbamazepine and its epoxide Gauchel, G.; Gauchel, FD. and Birkofer, L.: A micromethod for
metabolite in man after single and multiple doses. European the determination of carbamazepine in blood by high speed li-
Journal of Clinical Pharmacology 8: 337-341 (I 975). quid chromatography. Zeitung der klinische Chemie und
Ekbom, K.: Carbamazepine. a new symptomatic treatment for the klinische Biochemie II: 459-460 (1973).
paraesthesiae associated with Lhermitte's sign. Zeitschrift fur Gerardin, A. and Hirtz, J.: The quantitative assay of car-
Neurologie 200: 341-344 (1971). bamazepine in biological material and its application to basic
Clinical Pharmacokinetics of Carbamazepine 142
pharmacokinetic studies; in Birkmayer (Ed) Epileptic Seizures Kuroiwa, Y. and Shibasaki. H.: Carbamazepine for tonic seizure
- Behaviour - Pain, p, 151-164 (Huber, Berne 1976). in multiple sclerosis. Lancet I: 116 (\967).
Glatt, H.R.; Oesch, F.; Frigerio, A. and Garattini, S.: Epoxides Kutt, H.: Carbamazepine: Chemistry and methods of determina-
metabolically produced from some known carcinogens and tions; in Penry and Daly (Eds), Advances in Neurology, vol.
from some clinically used drugs: I. Differences in II, p.249-260 (Raven Press, New York 1975).
mutagenicity. International Journal of Cancer 16: 787-797 Kutt. H. and Penry, J.K.: Usefulness of blood levels of antiepilep-
(1975). tic drugs. Archives of Neurology 31: 283-288 (1974).
Goenechea, S. and Hecke-Seibicke, E.: Beitrag zum Stoffwechsel Levy, R.H.; Lockard, J.S.; Green, J.R.; Friel, P. and Martis, L.:
von Carbamazepine. Zeitung der klinische Chemie und Pharmacokinetics of carbamazepine in monkeys following in-
klinische Biochemie 10: 112-113 (1972). travenous and oral administration. Journal of Pharmaceutical
Gruska, H.; Beyer, K.-H.; Kubicki, S. and Schneider, H.: Klinik, Sciences 64: 302-307 (I 975a).
Toxikologie und Therapie einer schweren Carbamazepine- Levy. R.H.; Pitlick. W.H.; Troupin, A.S.; Green, J.R. and Neal,
Vergiftung. Archiv der Toxikologie 27: 193-203 (1971). J.M.: Pharmacokinetics of carbamazepine in normal man.
Hansen, J.M.; Siersbaek-Nielsen, K. and Skovsted, L.: Car- Clinical Pharmacology and Therapeutics 17: 657 -668 (197 5b).
bamazepine-induced acceleration of diphenylhydantoin and McAuliffe, J.J.; Sherwin, A.L.; Leppik, I.E.; Fayle, S.A. and Pip-
warfarin metabolism in man. Clinical Pharmacology and penger, C.E.: Salivary levels of anticonvulsants: A practical
Therapeutics 12: 539-543 (1971). approach to drug monitoring. Neurology 27: 409-413 (1977).
Hempel, E. and Klinger, W.: Drug stimulated biotransformation Meinardi, H.: Other antiepileptic drugs - carbamazepine; in
of hormonal steroid contraceptives: clinical implications. Woodbury, Penry and Schmidt (Eds) Antiepileptic Drugs.
Drugs 12: 442-448 (1976). p.487-496 (Raven Press, New York 1972).
Hooper, W.O.; Dubetz, D.K.; Bochner, F.; Cother, L.M.; Smith, Monaco, F; Riccio, A.; Benna, P.; Covacich, A.; Durelli, L.; Fan-
G.A.; Eadie, MJ. and Tyrer, J.H.: Plasma protein binding of tini, M.; Furlan. P.M.; Gilli, M.; Mutani, R.; Troni, W.; Ger-
carbamaiepine. Clinical Pharmacology and Therapeutics 17: na, M. and Morselli, P.L.: Further observation on car-
433-440 (1975). bamazepine plasma levels in epileptic patients - Relationships
Hooper, W.O.; Dubetz. O.K.; Eadie, MJ. and Tyrer, J.H.: with therapeutic and side effects. Neurology 26: 936-943
Preliminary observations on the clinical pharmacology of car- (I 976).
bamazepine (Tegretot). Proceedings of the Australian Associ- Morselli, P.L.: Carbamazepine: Absorption, distribution, and ex-
ation of Neurologists II: 189-198 (1974). cretion; in Penry and Daly (Eds) Advances in Neurology, vol.
Hvidberg, E.F. and Dam, M.: Clinical pharmacokinetics of anti- 11, p.279-293 (Raven Press. New York 1975).
convulsants. Clinical Pharmacokinetics I: 161-188 (1976). Morselli, P.L.; Baruzzi, A.; Gerna, M.; Bossi, L. and Porta, M.:
Johannessen, S.I.; Barruzzi. A.; Gomeni, R.; Strandjord, R.E. and Carbamazepine and carbamazepine-IO, I I-epoxide concentra-
_MorseUi, P.L.: further observations on carbam~t!pine and tigns in .h,uma.n brain. British Journal of Clinical Phar-
carbamazepine-I 0, II-epoxide kinetics in epileptic patients; in macology 4: 535-540 (1977).
Gardner-Thorpe, Janz, Meinardi and Pippenger (Eds) Anti- Morselli. P.L.; Biandrate. P.; Frigerio, A.; Gerna, M. and Tog-
epileptic Drug Monitoring, p. I 10-127 (Pitman Press, Avon noni, G.: Gas chromatographic determination of car-
1977). bamazepine and carbamazepine-I 0, II-epoxide in human body
Johannessen, S.l.; Gerna, M.; Bakke, J.; Strandjord, R.E. and fluids; in Meijer, Meinardi, Gardner-Thorpe and van der Klei-
MorseUi, P.L.: CSF concentrations and serum protein binding jn (Eds) Methods of Analysis of Anti-epileptic Drugs,
of carbamazepine and carbamazepine-I O,II-epoxide in epilep- p.169-175 (Excerpta Medica, Amsterdam 1973).
tic patients. British Journal of Clinical Pharmacology 3: Morselli, P.L.; Bossi, L. and Gerna, M.: Pharmacokinetic studies
575-582 (1976). with carbamazepine in epileptic patients; in Birkmayer (Ed)
Johannessen, S.1. and Strandjord, R.E.: The concentration of car- Epileptic Seizures - Behaviour - Pain, p.141-150 (Huber,
bamazepine (Tegretot) in serum and in cerebrospinal fluid in Berne 1976).
patients with epilepsy. Acta Neurologica Scandinavica 48 Morselli, P.L.; Gerna, M. and Garattini, S.: Carbamazepine
(Suppl. 51): 445-446 (1972). plasma and tissue levels in the rat. Biochemical Pharmacology
Johannessen, S.1. and Strandjord. R.E.: The influence of 20: 2043-2047 (\ 97 1).
phenobarbitone and phenytoin on carbamazepine serum levels; Murselli, P.L.; Gerna, M.; de Maio, D.; Zanda. G.; Viani, F. and
in Schneider, Janz, Gardner-Thorpe, Meinardi and Sherwin Garattini, S.: Pharmacokinetic studies on carbamazepine in
(Eds) Clinical Pharmacology of Anti-epileptic Drugs, volunteers and in epileptic patients, in Schneider, Janz,
p.201-205 (Springer-Verlag, Berlin 1975). Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Phar-
Kupferberg. HJ.: GLC determination of carbamazepine in macology of Anti-epileptic Drugs, p.166-179 (Springer-
plasma. J. Pharm. Sci. 61: 284-286 (1972). Verlag, Berlin 1975).
Clinical Pharmacokinetics of Carbamazepine 143
Neuvonen, PJ.; Pentilla, 0.; Lehtovaara, R. and Aho, K.: Effect Raskin, N.H.; Levinson, S.A.; Hoffman, P.M.; Pickett, J.B.E. and
of antiepileptic drugs on the elimination of various tetracycline Fields, H.L.: Postsympathectomy neuralgia: Amelioration
derivatives. European Journal of Clinical Pharmacology 9: with diphenylhydantoin and carbamazepine. American Jour-
147-154 (1975). nal of Surgery 128: 75-78 (I 974).
Oesch, F.: Mammalian epoxide hydrases: inducible enzymes Rasmussen, P. and Riishede, J.: Facial pain treated with car-
catalysing the inactivation of carcinogenic and cytotoxic bamazepine (Tegretol <11». Acta Neurologia Scandinavica 46:
metabolites derived from aromatic and olefmic compounds. 385-408 (I 970).
Xenobiotica 3: 305-340 (1973). Rawlins, M.D.; Collste, P.; Bertilsson, L. and Palmer, L.: Dis-
Palmer, L.; Bertilsson, L.; Collste, P. and Rawlins, M.: Quantita- tribution and elimination kinetics of carbamazepine in man.
tive determination of carbamazepine in plasma by mass frag- European Journal of Clinical Pharmacology 8: 91-96 (1975).
mentography. Clinical Pharmacology and Therapeutics 14: Richens, A. and Dunlop, F.: Serum-phenytoin levels in the man-
827-832 (1973). agement of epilepsy. Lancet 2: 247-248 (1975).
Parsonage, M.: Treatment with carbamazepine: Adults; ill Penry Schneider, H.: Carbamazepine: An attempt to correlate serum
and Daly (Eds) Advances in Neurology, vol.ll, p.221-234 levels with anti-epileptic and side effects; in Schneider, Janz,
(Raven Press, New York 1975). Gardner-Thorpe, Meinardi and Sherwin (Eds) Clinical Phar-
Penttila, 0.; Neuvonen, PJ.; Aho, K. and Lehtovaara, R.: Interac- macology of Anti-epileptic Drugs, p.151-1 S8 (Springer-
tion between doxycycline and some antiepileptic drugs. Britist Verlag, Berlin 1975a).
Medical Journal 2: 470-472 (1974). Schneider, H.: Carbamazepine: The influence of other antiepileptic
Pippenger, C.E.; Paris-Kutt, H.; Penry, J.K. and Daly, D.: Profi- drugs on its serum level; in Schneider, Janz, Gardner-Thorpe,
ciency testing in determinations of antiepileptic drugs. Journal Meinardi and Sherwin (Eds) Clinical Pharmacology of Anti-
of Analytical Toxicology I: I 18-122 (1977). epileptic Drugs, p.189-195 (Springer-Verlag, Berlin I 975b).
Pitlick, W.H.; Levy, R.H.; Troupin, A.S. and Green, J.R.: Phar- Simonsen, N.; Olsen, P.Z.; Kuhl, V.; Lund, M. and Wendelboe,
macokinetic model to describe self induced decreases in steady- J.: A comparative controlled study between carbamazepine
state concentrations of carbamazepine. Journal of Phar- and diphenylhydantoin in psychomotor epilepsy. Epilepsia I 7:
maceutical Sciences 65: 462-463 (1976). 169-176 (I 976).
Pynnonen, S. and Sillanpaa, M.: Carbamazepine in mother's milk. Strandjord, R.E. and Johannessen, S.l.: A preliminary study of
Lancet 2: 563 (1975). serum carbamazepine levels in healthy subjects and in patients
Pynnonen, S.; Sillanpaa, M.; Frey, H. and Iisa1o, E.: Car- with epilepsy; in Schneider, Janz, Gardner-Thorpe, Meinardi
bamazepine and its 10, II-epoxide in children and adults with and Sherwin (Eds) Clinical Pharmacology of Anti-epileptic
epilepsy. European Journal of Clinical Pharmacology II: Drugs, p.181-188 (Springer-Verlag, Berlin 1975).
129-133 (1977). Troupin, A.S.; Green, J.R. and Halpern, L.M.: Carbamazepine
Rane, A.; Bertilsson, L. and Palmer, L.: Disposition of placentally (Tegretol) as an anticonvulsant - A controlled double-blind
transferred carbamazepine (Tegretol <11» in the newborn. Euro- comparison with diphenylhydantoin (Dilantin). Acta
pean Journal of Clinical Pharmacology 8: 283-284 (I 975). Neurologica Scandinavica, Supp\. 60: 13-26 (I975).
Rane, A.; Hojer, B. and Wilson, J.T.: Kinetics of carbamazepine Wagner, J.G.; Northam, J.l.; Alway, C.D. and Carpenter, O.S.:
and its 10, II-epoxide metabolite in children. Clinical Phar- Blood levels of drug at the equilibrium state after multiple dos-
macology and Therapeutics 19: 276-283 (I 976). ing. Nature 207: 1301-1302 (I 965).
Rane, A.; Shand, D.G. and Wilkinson, G.R.: Disposition of car-
bamazepine and its 10, II-epoxide metabolite in the isolated
perfused rat liver. Drug Metabolism and Disposition 5: Author's address: Dr Leif Bertilsson. Department of Clinical
179-194 (I 977). Pharmacology, Huddinge Hospital, S-141 86 Huddinge (Sweden).