154

Epilepsy Res., 9 (1991) 154- 159 Elsevier

EPIRES 00425

Comparison of valproate concentrations in human plasma, CSF and brain tissue after administration of different formulations of valproate or valpromide

Heinz Gregor Wieser Department of Neurology, University Hospital Ziirich, CH-8091 Ziirich (Switzerland) (Received 28 February 1991; revision received 20 May 1991; accepted 31 May 1991) Key words: Valproate; Valpromide; Brain tissue concentration; CSF concentration

The concentration of valproate was measured in plasma, CSF and brain tissue of patients who underwent resective surgical treatment because of severe temporal lobe epilepsyafter pretreatment with either a sustained release formulation of valproate (Depakine Chrono®; 5 patients), the conventional formulation of vaiproate (Depakine®;6 patients) or valpromide (Depamide®;2 patients). With a mean serum value for all 13 patients of 32.3 #g/g valproate, the mean brain/serum ratio was 15.1% (SD 6.1%). The valproate concentration of the hippocampus was significantlyhigher than that of the amygdaia and patients who had the sustained release formulation had significantly higher valproate concentration in the CSF and in the hippocampal formation than those patients who had the conventional valproate. Since a few patients had tumors, whereas others had varying degrees of gliosis, it cannot be ruled out that these differences are the result of different histopathological conditions with related differences in blood-brain barrier functions.

INTRODUCTION Valproic acid (VPA; di-n-propylacetic acid; Depakine®) is an antiepileptic of major importance. Its role in other neurological and psychiatric disorders, however, remains to be defined. It has been marketed since 1967 as a sodium salt. Since 1987 a sustained release formulation (Depakine Chrono ®) has been available in Switzerland. Valpromide (VPD; dipropylacetamide; Depamide ®) is a primary amide of valproic acid. It is a prodrug of V P A and is biotransformed into valproic acid almost completely after oral or intravenous administration with a half-life of 0.84 +_ 0.33 h 2.

Correspondence to: Prof. Dr. H.G. Wieser, Neurology Department, University Hospital, Frauenklinikstr. 26, CH-8091 Ziirich, Switzerland. Tel. 0041-1-255 55 30; Fax 0041-1-255 44 29.

Like VPA, VPD possesses antiepileptic properties. VPD is also said to have antipsychotic activity I 1. From animal studies it is known that peak brain concentrations of valproate, which occur at about the same time as peak serum levels, are much lower. The brain levels of drug parallel those in plasma. In the mouse they are about 20% of serum levels3. There are, however, marked differences between different species, e.g., a given dose (mg/kg p.o.) produces lower serum and brain levels in rats than in mice. Thirty minutes after 250 m g / k g valproate p.o. the brain concentration in the mouse is 60 and that in the rat is only 15/~g Eq 14C-valproate/g tissue, and 120 min after 250 m g / k g valproate p.o. the brain concentration in the mouse is 49 and that in the rat is 14 ~g eq 14Cvalproate/g tissue 1,3. Furthermore, valproate does not disperse equally to all brain regions, but, according to Ciesielski et al. 4, it appears to accumulate preferen-

0920-1211/91/$03.50 © 1991 Elsevier Science Publishers B.V. All rights reserved

155 Phase II 3 months

Phase I 3 months D Conventional Depakine

R or Depakine Chrono

Washout

900 mg/day

U or Depamide

G S Phenytoin 300 mglday

CSF lumbar cisternal

X

X

Electroencephalogram

X

X

X

X

X

X

I

Serum concentration of

Hematological tests

VPA tissue concentration of:

XXXXXXXXXXXXXXX(XXXXXXXXXXXXXX: XXXXXXXXXXXXXXX

Clinical-epileptological evaluation

VPA and phenytoin

Surgery with:

XXXXXXXXXXXXXXX

XXXXXXXXXXXXXXX

900mg/day

Phase III 3 months

I

Blood Brain tissue amygdala hippocampus neocortex tumor (if present)

Fig. I. Schematic diagram of the study design. The examinations performed at each stage are marked with an x.

tially in brain areas with a high GABA transaminase activity. The plasma protein binding also differs depending on the species. At the blood levels usually found during therapy in man, about 90070 of valproate is bound to plasma proteins, mainly to albumin 6. The plasma protein binding is 79070 in dogs, 63070 in rats, and 12070 in mice. These differences might explain the large differences in half-life of valproate observed in different species. In rats and dogs plasma valproate levels fall in a bi-exponential manner, and in the rhesus monkey in a complicated tri-exponential fashion 5,1°. The halflives of the 3 phases have been calculated to be ct 12 min, ~ 102 min, -y 11.3 h 13. Since only the free form of valproate enters the brain, several factors, which can influence the percentage of unbound drug, have to be considered. Such factors are, for example, reduced albumin levels and substances which compete for the same protein binding sites. The latter include fatty acids, bilirubin, uric acid, and drugs such as aspirin, clofibrate, phenylbutazone, and anticonvulsant drugs such as diphenylhydantoin. The available human data suggest that the cerebrospinal fluid (CSF) in man contains valproate levels similar to the free blood level 7,9,2°. To the best of my knowledge human brain concentrations of valproate have been reported only by Vajda et al. 14 in 9 patients treated with valproate for 3 days before surgery. According to those authors 14 levels in brain

tissue varied between 6.8070 and 27.9070 of plasma levels. SUBJECTS AND STUDY DESIGN Thirteen patients (7 males and 6 females) between the ages of 18 and 43 years (mean age 30 +_ 8 years) were included in this study. Inclusion criteria were as follows: (1) willingness and competence to give informed consent; (2) patients suffering from drug-resistant partial epilepsy with complex partial seizures (cps) and therefore undergoing presurgical evaluation and operative therapy according to our established criteria formulated elsewhere]5; (3) no drugs other than phenytoin as comedication to VPA or VPD; (4) no evidence of noncompliance. Patients were excluded from this study for the following reasons: (1) any serious illness other than epilepsy; (2) any abnormalities of hepatic function other than those attributable to liver enzyme induction by antiepileptic drugs; (3) pregnancy, lactation, or risk of conception (women not using an intrauterine device). All patients of this study suffered from cps, but secondary generalized seizures have occurred in all patients. Presurgical evaluation included SEEG (stereoelectroencephalography 16) in 2 patients (Nos. 6, 13), SEEG in combination with foramen ovale electrode recording (FO 18) in one (No. 1), and FO in 9 patients

156 (Nos. 2, 3, 4, 5, 7, 8, 10, 11, 12). Patient No. 9 had no invasive presurgical seizure monitoring. The study design was a p p r o v e d by the Institutional Review Board. According to the study protocol (see Fig. 1) each patient was closely followed for at least 1 year. Clinical, E E G and l a b o r a t o r y examinations were p e r f o r m e d as indicated in Fig. 1. All patients had a 3month pretrial observation period and entered phase I with a m o n o t h e r a p y o f 300 m g / d a y phenytoin. Phase II served as a wash-out period for V P A or VPD. During phases I and III, in addition to phenytoin 300 m g / d a i l y , 6 patients took conventional Depakine; 5 patients received Depakine Chrono; and 2 patients had Depamide. The dosage was 3 times daily 300 mg (see Table I). A t the time o f the operation the plasma, lumbar a n d / o r cisternal C S F (5 cc), and brain tissue were collected as simultaneously as possible, all within half an hour. The time lag between the last dose of valproate or valpromide and sample collection was 14 (SD 0.7) h. If possible, a m y g d a l a and hippocampus and temporal neocortex, and - i f a t u m o r was present - tumor tissue were stored separately. Plasma, CSF, and brain tissue samples were immediately frozen at - 70°C. A 5-cc portion o f the same plasma sample was sent to the Hospital Laboratories for examination o f the serum concentrations o f V P A and phenytoin. Collected tissue samples were analyzed by Sanofi Laboratories, MontpeUier.

In all patients anesthesists followed the same principles. Anesthesia was initiated with 0 . 5 - 1 . 5 ml T h a l a m o n a l ® (1 ml contains 0.05 mg fentanyl and 2.5 mg droperidol) and was continued with fentanyl, droperidol, Pavulon ® (pancuronium bromide) and a nitrous oxide/O2-halothane mixture. A m y g d a l o h i p p o c a m p e c t o m y 17,21 was performed in all patients: on the right in 8 patients (Nos. 1, 2, 4, 5, 7, 8, 12, 13) and on the left in 5 (Nos. 3, 6, 9, 10, 11). Histological examination o f resected tissue revealed tumors or a h a m a r t o m a in 4 patients and h i p p o c a m p a l gliosis in all 9 other patients, the degree varying between slight and severe (see Table I). RESULTS With regard to the effect o f the valproate or valpromide treatment on the seizure frequency, there was only a slight decrease o f the mean seizure frequency (per month) during phase I when c o m p a r e d to the pretreatment period (17.6, SD 7.1 versus 19.3, SD 7.8) and likewise during phase III when c o m p a r e d to phase II (16.2, SD 5.2 versus 18.7, SD 5.4). Despite the fact that in each phase the first 2 weeks were ignored in order to take into account possible withdrawal effects during phase II, these differences did not reach statistical (F-test) significance. Table I shows the individual results and Table II

TABLE I I n d i v i d u a l V P A c o n c e n t r a t i o n s (l~g/g) in s e r u m , CSF, a n d brain tissue f o r the 13 p a t i e n t s

The patients received either Depakine (indicated by d), Depakine Chrono (c) or valpromide (v). Where possible, values for cisternal and lumbar CSF and for different anatomical brain compartments (amygdala and hippocampus) are given. The last row gives the histopathological findings. Patient 1d

2d

3d

4d

5d

6v

7v

8c

9d

10 c

11 c

12 c

13 c

Serum CSF Cisternal Lumbar Brain a Amygdala Hippocampus

19.7 1 9 . 4 30.4 40.2 45.6 2 1 . 1 4 1 . 1 2 7 . 5 1 3 . 0 26.2 4 7 . 5 4 1 . 7 46.5 3.05 3.6 3.3 2.9 7.8 3.0 4.4 3.65 1.0 8.2 6.1 7.5 6.0 3.05 2.8 2.9 7.8 3.0 4.9 4.3 8.2 3.8 3.6 3.8 3.9 3.0 1.0 8.4 7.5 6.0 1.4 3.85 5.45 5.98 5.7 2.96 3.6 3.7 2.05 7.7 11.03 4.4 3.9 1.4 1.5 7.1 4.15 5.85 3.1 0.7 5.9 7.05 3.3 1.8 1.4 6.2 4.3 5.3 5.9 2.7 1.6 9.1 14.9 8.1

Histopathologyb

sg

g

ham

g

g

g

g

g

g

a-c

o-d

g-g

sg

a 'Brain' values comprise all available data, i.e., amygdala, hippocampus a n d temporal neocortex (not indicated in this table). b g, gliosis; sg, severe g|iosis; ham, hamartoma; a-c, fibrillaryastrocytoma, WHO grade II; o-d, oligodendroglioma, WHO grade II; g-g, ganglioglioma of amygdala, WHO grade I.

157 TABLE II Tissue concentrations of valproate (#g/g) Tissue

Serum 1"- CSF L Brain ~ . Amygdala L.. Hippocampus

Table III shows the results in relation to the use of different formulations of valproate, and valpromide.

Mean

SD

n

Patients

32.30 4.65 4.75

11.52 2.15 2.43

13 13 13

1- 13 1 - 13 1 - 13

There are 2 significant findings. First, the valproate concentration in the CSF of patients on Depakine C h r o n o is significantly higher ( M a n n - W h i t n e y U-test,

3.595 5.95

2.253 10 3.85 10

1-7, 10, 11, 13 1-7, 10, 11, 13

6.27

2.96

3

3- 5

4.53 4.65

1.97 2.30

9 8

1, 3 - 8, 10, 11 2, 3, 7 - 9 , 11 - 13

Clot

14.72

5.80

10

Bone

4.58

Cortex .r:"CSF cisternal L-CSF lumbar

2.67 "

5

P = 0.026) than in patients with the convential formulation (Depakine). In the serum and the brain the same trends are apparent, but not statistically significant. Comparison of the concentration of valproate in the hippocampus of patients with Depakine Chrono shows the former to be significantly higher (MannWhitney U-test, P = 0.05) than that in patients who

4 - 13)

took the conventional formulation. The values for the b r a i n / s e r u m and the C S F / s e r u m ratios reflect these

3 - 5, 7, 8

findings (Table IV), but the differences are not

* NS; ** P = 0.049 (Wilcoxon, 1-tailed).

statistically significant. displays the mean tissue concentrations (in #g/g) of V P A in the serum, CSF and brain. In addition, the values for different brain compartments are shown, as well as for cisternal and lumbar CSF, for clot, and for bone. It should be observed that not all tissue compart-

DISCUSSION

ments could be analyzed in all patients. In 10 patients, however, amygdalar as well as hippocampal tissue concentrations could be compared and a significant difference was found. The surgical time lag between removal of the amygdala and the hippocampal formation cannot be the cause of this difference, because in

tients - collect CSF and plasma together with various b r a i n specimens in humans. Further advantages include the fact that the tissue triggering the habitual seizures can be analyzed (i.e., the primary epileptogenic area) and that tissue without gross pathology and with various epileptogenic abnormalities can be

this operation the amygdala is removed before the hippocampus. There was n o significant difference between CSF and brain tissue, nor between the cisternal

compared. Although there are 3 aspects of this study design, the m a i n goal of this study was to compare the valproate

and l u m b a r CSF.

concentrations in plasma, CSF and brain tissue. In ad-

Epilepsy surgery offers a unique possibility to conduct antiepileptic drug studies of the described type, because one can - without additional risks for the pa-

TABLE III Tissue concentrations of valproate (#g/g) and their break-down according to the drug formulation used All (n=13)

Depakine (n=6)

Depakine Chrono (n=5)

Depamide (n=2)

Serum Mean SD

32.3 11.52

i . "....... * 28.05 11.78

i 37.88 9.23

31.1 10.0

CSF Mean SD

I 3.61 2.05

**

4.65 2.15

1 6.29 1.56

3.7 0.7

Brain Mean SD

I 4.07 1.80

*

4.75 2.43

I 6.153 2.845

3.28 0.32

* NS; ** P = 0.051 (2-tailed), 0.026 (1-tailed) (Mann-Whitney U-test).

158 TABLE IV Ratios o f concentrations o f valproate in different compartments and their breakdown according to the drug formulation used A II (n = 13)

Depakine (n = 6)

CSF/Serum Mean SD

I 0.128 0.045

*

0.146 0.059

Brain/Serum Mean SD

~ 0.146 0.041

*

0.151 0.061

Depakine Chrono (n = 5)

Depamide (n = 2)

1 0.177 0.071

0.125 0.018

I 0.170 0.080

0.114 0.026

* NS (Mann-Whitney U-test).

dition, where possible, cisternal and lumbar CSF were measured and removed brain tissue was compartmentalized by anatomical (amygdala versus hippocampus versus temporal neocortex) and histopathological criteria (macroscopically 'normal' versus 'tumoral' tissue). Our results confirm that the CSF in man contains VPA levels similar to the free blood levels, a finding which has been previously described by other authors. Furthermore, we found that the VPA levels in brain tissue varied between 7.1070 and 29.5 070of plasma levels, corresponding extremely well to the figures reported by Vajda et al. 14 (6.807o and 27.907o respectively). The most interesting finding is that different anatomical structures showed different valproate concentrations, i.e., the amygdala exhibited significantly lower values than the hippocampus. Unfortunately our small sample size does not allow us to correlate the brain/serum and CSF/serum ratios with the histopathological findings of the resected brain specimens. There are, however, a few interesting observations. For example, the highest brain/serum quotient (29.5070 versus a mean of 15.1 070)was found in patient No. 10, with a marked behavioral syndrome resembling schizophrenia-like psychosis. This patient also had the highest CSF/serum ratio, amounting to 3 1.3°70 (mean 14.6070). The highest brain tissue concentration of valproate was found in patient No. 11. It amounted to 14.9/~g/g in the hippocampal tissue in close vicinity to an oligodendroglioma WHO II. Since in the amygdala of this patient the valproate concentration was only 3.3/,g/g, this patient also had the second largest difference between amygdaiar and hippocampal valproate levels. The largest difference between amygdalar and hippocampal concentrations

was found in patient No. 13 with a severe hippocampal gliosis. Moreover, both patients with a severe gliosis (Nos. 1 and 13, see Table I) had low brain/serum ratios (7.1°70 and 8.407o versus the group mean of 15.1070). These observations indicate that histopathology does affect local drug concentrations of the brain, probably due to an alteration of the blood-brain barrier and/or an alteration of the postulated 'active transport mechanisms '8. In another drug study, with a similar design but using flunarizine, differences between normal and tumorous tissue were even more striking. 19 The second aspect relates to the different formulations of valproate and valpromide. This comparison is of clinical relevance because some side effects of valproate, in particular the fetal antiepileptic drug syndrome 12, have been linked to excessive plasma peaks of VPA. Other dose-related side effects include anorexia, nausea, vomiting, and tremor, especially at the onset of therapy 12. Therefore, the slow release formulation, i.e., the protracted but not delayed absorption of valproate, might be advantageous, provided excessive peaking can be prevented and brain tissue concentrations of valproate are at least comparable to those obtained with the conventional formulation. This study documents that the slow release formulation guarantees brain tissue concentrations comparable to or even slightly higher than those obtained with the conventional formulation of VPA. In fact, the valproate values in the CSF and the hippocampus were significantly higher in those patients who received the slow release formulation than in the patients who had the conventional formulation of valproate. Our third conclusion concerns the efficacy of valproate in a homogeneous group of partial epilepsies with so-called drug-resistant complex partial seizures

159 o f proven mediobasal limbic origin, As is well known,

could not detect a significant seizure reduction when

the primary efficacy profile o f valproate was initially

valproate or valpromide was given together with

described as good for generalized seizure types, and in particular for absences and myclonias, and fair for

dition to the small sample size, all o f these patients had

generalized tonic-clonic and atonic seizures. In recent

partial epilepsy o f such severity that they underwent

phenytoin. It must however be emphasized that, in ad-

years, however, valproate has also been r e c o m m e n d e d

surgical treatment. If they had responded satisfactori-

as the antiepileptic drug o f second choice in symp-

ly to any available antiepileptic drug o f first choice,

tomatic partial epilepsies. With regard to efficacy, we

they would not have been included in this study.

REFERENCES 1 Aly, M.I. andAbdel-Latif, A.A.,Studiesondistributionand

2

3

4

5

6

7

8

9

10

11

metabolism of valproate in rat brain, liver and kidney, Neurochem. Res., 5 (1980) 1231 - 1242. Bialer, M., Rubinstein, A., Dubrovsky, J., Raz, I. and Abramsky, O., A comparative pharmacokinetic study of valpromide and valproic acid after intravenous administration in humans, Int. J. Pharmaceut., 23 (1985) 25 - 33. Chapman, A., Keane, P.E., Meldrum, B.S., Simiand, J. and Vernieres, J.C., Mechanism of anticonvulsant action of valproate, Prog. NeurobioL, 19 (1982) 315- 359. Ciesielski, L., MaRre, J., Cash, C. and Mandel, P., Regional distribution and effects of cerebral mitochondrial respiration of the anticonvulsive drug n-propylacetate, Biochem. Pharmacol., 24 (1975) 1055 - 1058. Dickinson, R., Harland, R. and llias, A., Dispositions of valproic acid in the rat - dose dependent metabolism distribution - enterohepatic circulation and choleretic effect, J. Pharmacol. Exp. Ther., 221 (1979) 583 - 595. Gugler, R. and Mueller, G., Plasma protein binding of valproic acid in healthy subjects and in patients with renal disease, Br. J. Clin. Pharmacol., 5 (1978)441- 446. Kupferberg, H.J., Sodium valproate. In" G.H. Glaser, J.K. Penry and D.M. Woodbury (Eds.), Antiepileptic Drugs: Mechanisms ofAction, Raven Press, New York, 1980, pp. 643 - 654. L6scher, W. and Eisenwein, H., Pharmacokinetics of sodium valproate in dog and mouse, Arzneimittel-Forsch., 28 (1978) 782 - 787. Meijer, J.W.A. and Hessing-Brand, L., Determination of lower fatty acid particularly the antiepileptic dipropylacetic acid in biological materials by means of micro diffusion and gas chromatography, Clin. Chim. Acta, 43 (1973) 215 - 222. Reckers-Ketting, J.J., Van Der Kleijn, E. and Loliveld, B.A., Pharmacokinetics of di-propyl-acetate (depakine) in different dosage forms in man and rhesus monkeys, Pharm. Weekbl., 110 (1975) 1232- 1236. Reynolds, J.E.F., Martindale, The Extra Pharmacopoeia, 28th Edn., Pharmaceutical Press, London, 1982, 1250 pp.

12 Schmidt, D., Pharmakotherapie von Epilepsien - Aktuelle Fragen und Kontroversen, Fortschr. Neurol., 51 (1983) 363 - 386. 13 Schobben, F., Vree, T., Van Der Kleijn, E., Clafssens, R. and Renier, O., Metabolism of valproic acid in monkeys and man. In: S.I. Johannessen, P.L. Morselli, C.E. Pippinger, A. Richens, D. Schmidt and H. Meinardi (Eds.), Antiepileptic Therapy: Advances in Drug Monitoring, Raven Press, New York, 1985, pp. 9 7 - 102. 14 Vajda, F.J.E., Donnan, G.A., Phillips, T. and Bladin, P.F., Human brain, plasma and cerebrospinai fluid concentration of sodium valproate after 72 hr of therapy, Neurology, 31 (1981) 486-487. 15 Wieser, H.G., Selectiveamygdalohippocampectomy: indications, investigative technique and results. In: L. Symon et al. (Eds.), Advances and Technical Standards in Neurosurgery, Vol. 13, Springer, Vienna, 1986, pp. 39- 133. 16 Wieser, H.G., Stereo-electroencephalography. In: H.G. Wieser and C.E. Elger (Eds.), Presurgical Evaluation of Epileptics, Springer, Berlin, 1987, pp. 192- 204. 17 Wieser, H.G., Selective amygdalo-hippocampectomy for temporal lobe epilpeys, Epilepsia, 29 (Suppl. 2) (1988) 100-113. 18 Wieser, H.G. and Moser, S., Improved multipolar foramen ovale electrode monitoring, J. Epilepsy, 1 (1988) 13- 22. 19 Wieser, H.G., Stodieck, S.R.G. and Elger, C.E., Chronic and add-on administration of flunarizine in patients with drug-resistant partial seizures: effect on seizures, surface and depth EEG, and cognitive performance. In: E.-J. Speckmann, H. Schulze and J. Walden (Eds.), Epilepsy and Calcium, Urban and Schwarzenberg, Munich, 1986, pp. 417-445. 20 Wulff, K., Flachs, H., Wurtz-Jorgensen, A. and Gram, L., Clinical pharmacological aspects of valproate sodium, Epilepsia, 18 (1977) 149- 157. 21 Yasargil, M.G. and Wieser, H.G., Selective microsurgical resections. In: H.G. Wieser and C.E. Elger (Eds.), PresurgicalEvaluation of Epileptics, Springer, Berlin, 1987, pp. 352- 360.