- Email: [email protected]
Variability and clinical relevance of the interaction between sodium valproate and carbamazepine in epileptic patients
193
Epilepsy Res., 11(1992) 193-198 Elsevier EPIRES 00460
Variability and clinical relevance of the interaction between sodium valproate and carbamazepine in epileptic patients
P. J. W. McKee, J. Blacklaw, E. Butler, R.A. Gillham” and M. J. Brodie Epilepsy Research Unit, University Department of Medicine and Therapeutics, Western lnf%nuuy and “Department of Clinical Psychology, Institute of Neurological Sciences, Southern General Hospital, Glasgow (UK)
(Received 2 December 1991; accepted 20 December1991) Key words: Carbamaxepine;
Sodium valproate; Pharmacokinetics;
Drug interaction; Cognitive function
Twenty-four epileptic patients (16 females, 8 males; aged 13-62 years) were studied before and after the addition of sodium valproate (VPA) 500 mg twice daily for 5 days. All had been established previously on carbamaxepine (CBZ) as monotherapy (300-1600 mg daily in divided doses). Sixteen of these patients undertook a battery of cognitive function tests before and after VPA introduction. VPA had no effect on total or free CBZ concentrations. However, median concentrations of the active metabohte, CBZ 10,ll epoxide (CBZ-E), were significantly increased (CBZ-E before VPA 1.3 mg/l, after VPA 2.1 mg/l, P < 0.01). The median rise was 25%, although the extent of the interaction ranged from a 25% decrease to an increase of 123% in CBZ-E concentrations. This was related to the marked inter-individual variation in circulating VPA (mean 25-69 mg/l), as CBZ-E concentrations correlated significantly with total (r = 0.5, P < 0.05,95% CI 0 to +0.08) and free (r = 0.7, P < 0.001,95% CI +0.09 to +0.25) VPA levels in individual patients. Although uncontrolled, no deterioration in performance of any of the cognitive function tests was observed following the addition of VPA. This study does not support immediate clinical relevance for this drug interaction between VPA and CBZ.
INTRODUCTION Carbamazepine (CBZ) and sodium valproate (VPA) are first-line antiepileptic drugs used singly in the treatment of patients with partial and generalised tonic-clonic seizuresiY2 and in combination in the management of refractory epilepsg-6. CBZ 10,ll epoxide (CBZ-E), a major metabolite of CBZ, possesses anticonvulsant activity7 and has been implicated in some of the side-effects associated with the drug8-i4. Correspondence to: Dr Martin J. Brodie, Epilepsy Research Unit, Department of Medicine and Therapeutics, Western Infirmary, Glasgow Gil 6NT, UK.
Higher CBZ-E levels have been recorded in epileptic patients taking CBZ in combination with VPA than in those treated with CBZ monotherapy”@. This is a consequence of inhibition by VPA of epoxide hydrolase, the enzyme responsible for the breakdown of CBZ-E to the inert dihydrodiol’7-‘9. In a previous placebo-controlled evaluation in healthy volunteers, the addition of VPA was associated with a 52% increase in CBZE levels and some deterioration in psychomotor performance20. Other workers have demonstrated considerable variation in the extent of this interaction in small numbers of epileptic patients using different VPA dosage schedules21>22. Although now well delineated, the clinical rele-
0920-1211/92/$05.00 tQ 1992 Elsevier Science Publishers B.V. All rights reserved
194
Vance of this interaction between VPA and CBZ remains to be established. We undertook to monitor the effect on CBZ and CBZ-E concentrations and cognitive function of the introduction of a standard dose of VPA in epileptic patients established on CBZ as monotherapy, in whom such a change in management was clinically indicated. METHODS Patients Twenty-four patients (8 males, 16 females; aged 13-62 years) referred to the epilepsy clinic at the Western Infirmary in Glasgow were studied (Table I). All had been established on CBZ monotherapy and had taken the drug in the same dose and timing for at least 2 months prior to entry. To-
TABLE
I
Clinical characteristics of 24 Sex Age (years)
epilepticpatients
Weight Seizure
(kg)
we
Carbama-
Dosage
zepine
interval
dose (mglday)
(h)
51
F
91
PC
800
35
F
58
MYO/GTCS
800
6
24
F
61
PSG
800
12
12
16
M
60
PSG
400
6
28
F
52
PSG
400
12
58
F
61
PC
24
F
57
GTCS
400
12
1000
38
M
77
12
PSG
800
25
F
6
55
PSG
600
32
8
F
60
GTCS
300
12
23
F
58
MYo/pc
800
12
33
F
58
MYO/GTCS
900
12
35
M
74
PSG
1000
12
62
F
67
GTCS
27
M
74
PSG
26
F
65
MYO/T’SG
800
12
51
F
67
GTCS
1200
12
53
F
65
58
F
64
PC GTCS
800 400
12 12
23
M
72
MYo/Pc
400
12
32
F
58
MYo/Pc
400
12
23
M
65
GTCS
1600
6
53
M
74
GTCS
1100
8
13
M
55
MYO/GTCS
1600
12 ____
800
8
1200
12
PC, partial complex seizures; GTCS, generalised tonic clonic seizures; MYO, myoclonic seizures; PSG, partial seizures with secondary generalisation.
tal daily dosage ranged from 300 to 1600 mg. Seventeen patients took CBZ twice daily, 3 patients were on a thrice daily regimen and 4 patients received the drug in four divided doses. The decision to introduce VPA was made on clinical grounds by the attending physician. In all patients the subsequent intention was to withdraw slowly the CBZ leaving the patient on VPA monotherapy. In 13 patients seizures were poorly controlled. Seven of these had previously undiagnosed myoclonic epilepsy. Mild but persistent side-effects were the reason for adding VPA in a further 6. The remaining 5 patients had suboptimal seizure control coupled with intermittent symptoms consistent with CBZ neurotoxicity. No other medication was taken in the preceding month nor during the short experimental period. The study had the approval of the Western Infirmary Ethical Committee and written informed consent was obtained from all participants. Protocol All patients were admitted to hospital several days before and throughout the period of study to ensure compliance with the treatment schedule. On the third day after admission, blood samples were taken every 2 h from an indwelling catheter inserted in a suitable peripheral vein immediately after the morning CBZ dose (09.00 h) and over one dosage interval. VPA 500 mg twice daily was introduced with the timing standardised at 09.00 h and 21.00 h each day. Venous blood sampling was repeated 5 days later. All patients tolerated the combination of CBZ and VPA and there were no withdrawals from the study. Cognitive function Sixteen patients performed a battery of cognitive function tests immediately before and 1 and 8 h after the morning CBZ dose during the day prior to and on the fifth day after the introduction of VPA. Some of these tests had been shown to be sensitive to the detrimental effects of CBZ and CBZ-El’. They included: (1) Decision time: time (ms) to respond to a light coming on by removing the finger from the base button on Leeds Psychomotor Tester. A mean of 30 tests was recorded.
195
(2) Movement time: time (ms) to move the finger from the base button to extinguish a light on Leeds Psychomotor Tester. A mean of 30 tests was recorded. (3) Finger tapping: number of taps (per minute) of the dominant index finger on a calculator button in constant addition mode. (4) Forward digit span: maximum number of digits the subject could recall following oral presentation. He or she was allowed two trials at each level and the task was discontinued when both were failed. (5) Backward digit span: maximum number of digits the subject could recall in reverse order immediately following oral presentation. He or she was allowed two trials at each level and the task was discontinued when both were failed. (6) Paired associate learning: number of trials to reach the criterion of three correct in learning unrelated word pairs. (7) Threshold detection: an array of small rectangles was displayed on a visual display unit. After a brief period of time, an extra rectangle was added. The subject was required to indicate which it was. The ‘threshold’ was the minimum time gap which the subject required to perceive that an extra rectangle has been added. (8) Sedation score: subjects rating of level of alertness using a lo-cm line, where 0 was ‘nearly asleep’ and 10 ‘wide awake’. (9) Side-effect score: the subject was presented with a standardised list of 15 symptoms (e.g., dizziness, diplopia). Each one was graded zero, one, two, three, depending on severity. Individual ratings were summated to produce a total score. TABLE
Analyses Blood samples were taken into heparinised tubes, centrifuged immediately and the plasma stored at -20°C for batch analysis. Total CBZ and VPA concentrations were obtained by enzyme immunoassay (Emit, Syva, Palo Alto, CA, USA) and free CBZ23 and VPAz4 were determined by equilibrium dialysis and enzyme immunoassay. CBZ-E concentrations were assayed by high-performance liquid chromatography as previously describedB. Areas under the concentration-time curve (AUC) for a dosage interval were calculated using the trapezoidal rule. Statistics were performed on antiepileptic drug concentrations using the Wilcoxon signed-rank test for matched pairs. Cognitive function data were examined by repeated-measures two-way analysis of variance, and differences between subgroups by Mann-Whitney U test for unpaired samples. The significance of correlations was explored using linear regression. RESULTS Pharmacokinetics No changes in total or free CBZ concentrations were found following the addition of VPA (Table II). However, CBZ-E concentrations were significantly increased for the whole patient group. This is illustrated for the 17 patients taking CBZ twice daily in Fig. 1. Overall, there was a median 25% increase in mean CBZ-E levels following VPA treatment. However, there was considerable inter-individual variation in the extent of this interaction, ranging from a 25% decrease to an in-
II
Median corb~~epine (‘CBZ) and carob~zep~~e l&U epoxide (CBZ-E) concen~~‘o~ in 24 ep~epticpact addition of sodium valproate 500 mg twice daily for 5 days Total CBZ
Free CBZ
Without VPA
With VPA
95% CI Without for di~ere~ce VPA
CBZ-E With VPA
95% CI for d#erence
Without VPA
With VPA
95% CI for di~~ence
2.1** 2.7.
+0.2 to +0.9
1.5**
+0.1 to +os
20***
+2
Mean (m&I)
8.4
8
-0.3 to +1.2
2
1.8
-0.1 to +0*3
1.3
Maximum (men)
9.4
8.8
-0.1 to +1.5
2.6
2.2
-0.1 to +0.4
1.8
Minimum (mg/l) AUC (me)
6.9 80
6.8 88
-0.6 to +0.9 -4.2 to + 12
1.6 22
1.6 21
-0.2 to +0.2 -0.5 to +4.0
1.1 15
*P<005~**P<001~‘**P<0.001. . f . t
before and after the
+0.1 to +1.0 to +9.5
196 Clrblmarrpina 10
10 -
11 epoxida a
conc~ntrrtionhg/ll
,.Ol 1
a-
T
Number Patients
of 6-
2.6 42.2 21.6 O1.4
1.0 J, 0
2
4
6
6
Hours after crrbamazapine
I 12
10
dose
-25 - 0
0 - 24
25.49
50-74
>75
Fig. 2. Percentage change in median carbamazepine 10,ll epoxide concentrations following the addition of sodium valproate 500 mg twice daily for 5 days in 24 epileptic patients receiving carbamazepine as monotherapy.
Fig. 1. Carbamazepine 10,ll epoxide concentrations over a dosage interval before and after the addition of sodium valproate 500 mg twice daily for 5 days in 17 epileptic patients taking carbamazepine in twice daily dosage.
Cognitive function
crease of 123% in median CBZ-E concentrations (Fig. 2). Mean VPA concentrations over a dosage interval varied widely between patients with a median total VPA of 44 mg/l (range 25-69 mg/l) and a free value of 4.3 mg/l (range 1.6-6.8 mg/l). Mean CBZ-E concentrations in individual patients correlated significantly with mean total (r = 0.5, P < 0.05, 95% CI 0 to +0.08) and free (r = 0.7, P < 0.001,95% CI +0.09 to +0.25) VPA levels.
The results of cognitive function testing in the 16 patients are shown in Table III. The addition of VPA was not associated with any deterioration in performance in any of the tests. However, mean threshold detection scores in the subgroup of patients (n = 7) showing a greater than 25% increase in median CBZ-E concentrations following the addition of VPA were higher than in those (n = 9) with little change (4.8 versus 1.7, P < 0.05). There were no significant differences in performance between these groups for the other tests in the battery (data not shown).
TABLE III Cognitive function tests before and I and 8 h afier the morning carbamazepine
dose with and without additional
sodium
valproate
I6 epileptic patients Test
Decision time (ms)* Movement time (ms)* Finger tapping (per 15 s) Forward digit span Backward digit span Verbal learninga Threshold detectiona (frame units) Sedation scorea (cm) Side-effect score=
Before valproate
After valproate Oh
Ih
8h
(0.3) (0.2) (18) (1.5) (1.6) (4)
0.60 (0.4) 0.31 (0.1) 76 (13) 6.2 (1.5) 4.1 (1.4) 10.8 (5)
0.56 (0.2)
0.57 (0.2) 0.34 (0.1) 75 (17) 5.9 (1.4) 4.7 (1.7) 10.1 (5)
3.2 (3.5) 2.9 (3.0) 9.4 (8)
3.1 (3.0) 2.0 (2.3) 7.2 (7)
3.2 (3.4) 1.8 (2.6) 8.5 (11)
Oh
lh
8h
0.52 (0.2) 0.31(0.1) 74 (20) 6.3 (1.3) 4.1 (1.2) 11.1(5)
0.54 (0.3) 0.32 (0.1) 71 (19) 6.8 (1.1) 4.3 (1.9) 10.2 (4)
0.62 0.34 73 6.3 4.3 9.9
3.3 (3.9) 3.3 (3.0) 13.2 (8)
3 (3.0) 2.1 (2.3) 9.8 (9)
All values are mean (SD). a Denotes test where low score is better than high score.
0.32 (0.1)
71 (17) 5.9 (1.8) 4.5 (1.6) 10.1 (6)
3.3
(3.6)
2.9 7.5
(2.9) (8)
in
197
DISCUSSION Reports of the effect of VPA on total CBZ concentrations are conflicting with a risez6,decrease” or no change%,” noted in various studies. Levy and his colleagues21 reported VPA-mediated inhibition of CBZ metabolism during intravenous infusions in rhesus monkeys. This lack of consensus may be partly explained by a small additional effect of VPA in displacing CBZ from binding sites on plasma proteins20*21,27, Higher free CBZ levels with prolongation of the elimination half-life have been observed in healthy volunteers taking both drugs ” . In the present study, however, in 24 epileptic patients we observed no significant alterations in total or free CBZ concentrations after 5 days’ treatment with a sizeable dose of VPA. The interaction between CBZ and VPA has generated some interest since increased levels of CBZ-E were first recorded in epileptic patients taking both drugs’5,‘6.The effect of VPA on CBZE metabolism is a consequence of inhibition of epoxide hydrolase, the enzyme responsible for its breakdown to the inert dihydrodio117-‘9.Although side-effects experienced by some patients on the combination of CBZ and VPA have been attributed to unsuspected high concentrations of CBZE ‘*,14,other workers have found no relationship between CBZ-E levels and toxic events28. In the only placebo-controlled evaluation, a mean 52% increase in CBZ-E levels was associated with some deterioration in psychomotor performance in healthy volunteers*‘. Considerable variation in the extent of this interaction has been demonstrated in small numbers of patients with epilepsy using different VPA doses. Pisani and colleagues= reported an increase in CBZ-E concentrations between 29% and 23g%, whereas Levy and coauthors*l observed a range of changes from a 14.7% decrease to an 148.4% increase. In our study most patients showed an elevation in CBZ-E concentrations during VPA co-administration. The median increase was 25%. However, there was considerable variation in the magnitude of this
effect, which varied from a 25% decrease to an increase of 123% in mean CBZ-E levels. Part of this va~abi~ty can be accounted for by the substantial differences in circulating VPA concentrations among patients taking the same dose. This may be, in part, due to a combination of concentrationdependent protein binding9 and differences in VPA clearance as a consequence of enzyme induction by CBZ30,3’.The observed correlation between total and free VPA and CBZ-E concentrations suggests that the extent of this effect depends on the VPA dosage. Accordingly, the higher the dose of both drugs, the more likely that this interaction will be clinically relevant. The relationship between seizures, anticonvulsant drugs, psychosocial difficulties and impairment of cognitive function is highly complex. We have shown in previous work that high CBZ-E levels were associated with poorer performance in our battery of cognitive function tests”. In this study we could demonstrate no deterioration in performance of any test after the addition of VPA. Although these observations were uncontrolled and may, therefore, be liable to bias, they do suggest that patients receiving CBZ as monotherapy tolerate the introduction of VPA at this dosage despite an increase, sometimes substantial, in CBZE concentrations. Exploration of the interaction between CBZ and VPA has demonstrated that, although pharmacokinetically significant, there is considerable variation in its magnitude. Although CBZ-E concentrations are not readily predictable in a given patient, it is unlikely that measurement of this metabolite will be helpful in this clinical setting. The adverse effects of this interaction is likely to be small in all but a few patients receiving large doses of both drugs.
Our grateful thanks go to Anne Somers for expert secretarial assistance.
198 REFERENCES 1 Editorial, Carbamazepine update. Lancer, ii (1989) 595-597. 2 Editorial, Sodium valproate, Lancer, ii (1988) 1229- 1231. 3 Aicardi, J., Clinical approach to the management of intractable epilepsy, Dev. Med. Child. Neural., 30 (1988) 429-440. 4 Brodie, 5
6 7
8
M.J., Established anticonvulsants and the treatment of refractory epilepsy, Lancer, 336 (1990) 3.50-354. Shorvon, S.D., Medical assessment and treatment of chronic epilepsy, Br. Med. J., 302 (1991) 363-366. Scheuer, M.L. and Pedley, R.A., The evaluation and treatment of seizures, N. Engf. J. Med., 233 (1991) 1468-1474. Tomson, T., Almkvist, O., Nilsson, B.Y., Svensson, J.O. and Bertilsson L., Carbamazepine IO,11 epoxide in epilepsy. A pilot study, Arch. Neurol., 47 (1990) 888-892. Lindhout, D., Hoppener, R.J.E.A. and Meinardi, H., Teratogenesis of antiepileptic drug combinations with special emphasis on epoxidation (of carbamazepine), Epilepsia, 25 (1984) 77-83.
9 Patsalos, P.N., Stephenson,
T.J., Krishna, S., Elyas, A.A., Lascelles, P.T. and Wiles, C.M., Side-effects induced by carbamazepine 10,ll epoxide, Lancet, ii (1984) 496. 10 Rosenfield, W.E., Holmes, G.B., Graves, N.M. and Peterson, A.L., Carbamazepine 10,lI epoxide serum concentration as a predictor of toxicity, Epilepsia, 28 (1987) 580-581. 11 Gillham, R., Williams, N., Wiedmann, K., Butler, E., Larkin, J.G. and Brodie, M.J., Concentration-effect relationships with carbamazepine and its epoxide on psychomotor and cognitive function in epileptic patients, J. Neurol. Neurosurg. Psychiatry, 51 (1988) 929-933. 12 McKee, P.J. W., Larkin, J.G. and Brodie, M.J., Acute psychosis with carbamazepine and sodium valproate, Lancer, i (1989) 167. 13 Kerr, M. and Levy, R.H., Inhibition of epoxide hydrolase by anticonvulsants and risk of teratogenicity, Lance& i (1989) 610-611. 14 Rambeck, B., Salke-Treumann, A., May, T.H. and Boenigk, H.E., Valproic acid induced carbamazepine 10,ll epoxide toxicity in children and adolescents, Eur. Neurol., 30 (1990) 79-83. 15 McKauge, L., Tyrer, J. and Eadie, M.J., Factors influencing simultaneous concentrations of carbamazepine and its epoxide in plasma, Ther. Drug Monit., 3 (1981) 63-70. 16 Brodie, M.J., Forrest, G. and Rapeport, W.G., Carbamazepine lo,11 epoxide concentrations in epileptics on carbamazepine alone and in combination with other anticonvulsants, Br. .I. Clin. Pharmacol., 16 (1983) 747-750. 17 Chang, S.L. and Levy, R.H., Inhibition of epoxidation of carbamazepine by valproic acid in the isolated perfused rat liver, J. Pharmacokinet. Biopharm., 13 (1985) 453-466. 18 Pisani, F., Caputo, M., Fazio, A., Oteri, G., RUSSO,M., Spina, E., Perucca, E. and Bertilsson, L.. Interaction of
carbamazepine 10,ll epoxide, an active metabohte of carbamazepine, with valproate: a pharmacokinetic study, Epilepsia, 31 (1990) 339-342. 19 Robbins, D.K., Wedlund, P. J., Kuhn, R., Baumann, R.J ,
Levy, R.H. and Chang, S.L., Inhibition of epoxide hydrolase by valproic acid in epileptic patients receiving carbamazepine, Br. J. Clin. Pharmacol., 29 (1990) 759-762, 20 Macphee, G.J.A., Mitchell, J., Wiseman, L., McLellan, A.R., Park, B.K., McInnes, G.T. and Brodie, M.J., Effect of sodium valproate on carbamazepine disposition and psychomotor profile in man, Br. J. Clin. Pharmacol., 25 (1988) 59-66. 21 Levy, R.H., Moreland,
T.A., Morselh, P.L., Guyot, M., Brachet-Liermain, A. and Loiseau, P., Carbamazepinel valproic acid interaction in man and rhesus monkey, Epi-
iepsia, 25 (1984) 338-345. 22 Pisani, F., Fazio, A., Otari, G., Ruello, C., Gitto, F., Rus-
so F. and Perucca, E., Sodium valproate and valpromide: differential interactions with carbamazepine in epileptic patients, Epifepsia, 27 (1986) 548-552. 23 Macphee, G.J.A., Goldie, C., Poulston, D., Potter, L.. Agnew, E., Laidlaw, J. and Brodie, M.J., Effect of carbamazepine on psychomotor performance in naive subjects, Eur. .I. Clin. Pharmacol.,
30 (1986) 37-42.
24 Rapeport,
W.G., Mendelow, A.D., French, G., Macpherson, P., Teasdale, E., Agnew, E., Thompson, G.G. and Brodie, M.J., Plasma protein-binding and CSF concentrations of valproic acid in man following acute oral dosing, Br. J. Clin. Pharmacol.,
16 (1983) 365-369.
25 Macphee,
26 27
28
29
30
G.J.A., Thompson, G.G., Scobie, G.. Agnew. E., Park, B.K., Murray, T., McCall, K.E.L. and Brodie, M.J., Effects of cimetidine on carbamazepine auto- and hetero-induction in man, Br. J. Clin. Pharmacol., 18 (1984) 411-419. Jeavons, P.M. and Clarke, J.E., Sodium valproate in the treatment of epilepsy, Br. Med. J., 2 (1974) 584-586. Mattson, G.F., Mattson, R.H. and Cramer, J.A., Interaction between valproic acid and carbamazepine: an in vitro study of protein binding, Ther. Drug Monir., 4 (1982) 181-184. Theodore, W .H., Narang, P.K., Holmes, M.D., Reeves. P. and Nice, F.J., Carbamazepine and its epoxide: relation of plasma levels to toxicity and seizure control, Ann. Neurol., 25 (1989) 194-196. Bowdle, T.A., Pate& I.H., Levy, R.H., and Wilensky, A.J. Valproic acid dosage and plasma protein binding and clearance, Clin. Pharmacol. Ther., 28 (1981) 485-492. Bowdle, T.A., Levy, R.H. and Cutler, R.E., Effect of carbamazepine on valproic acid kinetics in normal subjects, Clin. Pharmacol. Ther., 26 (1979) 629-634.
31 May T. and Rambeck, B., Serum concentration
acid: influence of dose and co-medication, Monk,
6 (1984) 334-343.
of valproic Ther.
Drug
Epilepsy Res., 11(1992) 193-198 Elsevier EPIRES 00460
Variability and clinical relevance of the interaction between sodium valproate and carbamazepine in epileptic patients
P. J. W. McKee, J. Blacklaw, E. Butler, R.A. Gillham” and M. J. Brodie Epilepsy Research Unit, University Department of Medicine and Therapeutics, Western lnf%nuuy and “Department of Clinical Psychology, Institute of Neurological Sciences, Southern General Hospital, Glasgow (UK)
(Received 2 December 1991; accepted 20 December1991) Key words: Carbamaxepine;
Sodium valproate; Pharmacokinetics;
Drug interaction; Cognitive function
Twenty-four epileptic patients (16 females, 8 males; aged 13-62 years) were studied before and after the addition of sodium valproate (VPA) 500 mg twice daily for 5 days. All had been established previously on carbamaxepine (CBZ) as monotherapy (300-1600 mg daily in divided doses). Sixteen of these patients undertook a battery of cognitive function tests before and after VPA introduction. VPA had no effect on total or free CBZ concentrations. However, median concentrations of the active metabohte, CBZ 10,ll epoxide (CBZ-E), were significantly increased (CBZ-E before VPA 1.3 mg/l, after VPA 2.1 mg/l, P < 0.01). The median rise was 25%, although the extent of the interaction ranged from a 25% decrease to an increase of 123% in CBZ-E concentrations. This was related to the marked inter-individual variation in circulating VPA (mean 25-69 mg/l), as CBZ-E concentrations correlated significantly with total (r = 0.5, P < 0.05,95% CI 0 to +0.08) and free (r = 0.7, P < 0.001,95% CI +0.09 to +0.25) VPA levels in individual patients. Although uncontrolled, no deterioration in performance of any of the cognitive function tests was observed following the addition of VPA. This study does not support immediate clinical relevance for this drug interaction between VPA and CBZ.
INTRODUCTION Carbamazepine (CBZ) and sodium valproate (VPA) are first-line antiepileptic drugs used singly in the treatment of patients with partial and generalised tonic-clonic seizuresiY2 and in combination in the management of refractory epilepsg-6. CBZ 10,ll epoxide (CBZ-E), a major metabolite of CBZ, possesses anticonvulsant activity7 and has been implicated in some of the side-effects associated with the drug8-i4. Correspondence to: Dr Martin J. Brodie, Epilepsy Research Unit, Department of Medicine and Therapeutics, Western Infirmary, Glasgow Gil 6NT, UK.
Higher CBZ-E levels have been recorded in epileptic patients taking CBZ in combination with VPA than in those treated with CBZ monotherapy”@. This is a consequence of inhibition by VPA of epoxide hydrolase, the enzyme responsible for the breakdown of CBZ-E to the inert dihydrodiol’7-‘9. In a previous placebo-controlled evaluation in healthy volunteers, the addition of VPA was associated with a 52% increase in CBZE levels and some deterioration in psychomotor performance20. Other workers have demonstrated considerable variation in the extent of this interaction in small numbers of epileptic patients using different VPA dosage schedules21>22. Although now well delineated, the clinical rele-
0920-1211/92/$05.00 tQ 1992 Elsevier Science Publishers B.V. All rights reserved
194
Vance of this interaction between VPA and CBZ remains to be established. We undertook to monitor the effect on CBZ and CBZ-E concentrations and cognitive function of the introduction of a standard dose of VPA in epileptic patients established on CBZ as monotherapy, in whom such a change in management was clinically indicated. METHODS Patients Twenty-four patients (8 males, 16 females; aged 13-62 years) referred to the epilepsy clinic at the Western Infirmary in Glasgow were studied (Table I). All had been established on CBZ monotherapy and had taken the drug in the same dose and timing for at least 2 months prior to entry. To-
TABLE
I
Clinical characteristics of 24 Sex Age (years)
epilepticpatients
Weight Seizure
(kg)
we
Carbama-
Dosage
zepine
interval
dose (mglday)
(h)
51
F
91
PC
800
35
F
58
MYO/GTCS
800
6
24
F
61
PSG
800
12
12
16
M
60
PSG
400
6
28
F
52
PSG
400
12
58
F
61
PC
24
F
57
GTCS
400
12
1000
38
M
77
12
PSG
800
25
F
6
55
PSG
600
32
8
F
60
GTCS
300
12
23
F
58
MYo/pc
800
12
33
F
58
MYO/GTCS
900
12
35
M
74
PSG
1000
12
62
F
67
GTCS
27
M
74
PSG
26
F
65
MYO/T’SG
800
12
51
F
67
GTCS
1200
12
53
F
65
58
F
64
PC GTCS
800 400
12 12
23
M
72
MYo/Pc
400
12
32
F
58
MYo/Pc
400
12
23
M
65
GTCS
1600
6
53
M
74
GTCS
1100
8
13
M
55
MYO/GTCS
1600
12 ____
800
8
1200
12
PC, partial complex seizures; GTCS, generalised tonic clonic seizures; MYO, myoclonic seizures; PSG, partial seizures with secondary generalisation.
tal daily dosage ranged from 300 to 1600 mg. Seventeen patients took CBZ twice daily, 3 patients were on a thrice daily regimen and 4 patients received the drug in four divided doses. The decision to introduce VPA was made on clinical grounds by the attending physician. In all patients the subsequent intention was to withdraw slowly the CBZ leaving the patient on VPA monotherapy. In 13 patients seizures were poorly controlled. Seven of these had previously undiagnosed myoclonic epilepsy. Mild but persistent side-effects were the reason for adding VPA in a further 6. The remaining 5 patients had suboptimal seizure control coupled with intermittent symptoms consistent with CBZ neurotoxicity. No other medication was taken in the preceding month nor during the short experimental period. The study had the approval of the Western Infirmary Ethical Committee and written informed consent was obtained from all participants. Protocol All patients were admitted to hospital several days before and throughout the period of study to ensure compliance with the treatment schedule. On the third day after admission, blood samples were taken every 2 h from an indwelling catheter inserted in a suitable peripheral vein immediately after the morning CBZ dose (09.00 h) and over one dosage interval. VPA 500 mg twice daily was introduced with the timing standardised at 09.00 h and 21.00 h each day. Venous blood sampling was repeated 5 days later. All patients tolerated the combination of CBZ and VPA and there were no withdrawals from the study. Cognitive function Sixteen patients performed a battery of cognitive function tests immediately before and 1 and 8 h after the morning CBZ dose during the day prior to and on the fifth day after the introduction of VPA. Some of these tests had been shown to be sensitive to the detrimental effects of CBZ and CBZ-El’. They included: (1) Decision time: time (ms) to respond to a light coming on by removing the finger from the base button on Leeds Psychomotor Tester. A mean of 30 tests was recorded.
195
(2) Movement time: time (ms) to move the finger from the base button to extinguish a light on Leeds Psychomotor Tester. A mean of 30 tests was recorded. (3) Finger tapping: number of taps (per minute) of the dominant index finger on a calculator button in constant addition mode. (4) Forward digit span: maximum number of digits the subject could recall following oral presentation. He or she was allowed two trials at each level and the task was discontinued when both were failed. (5) Backward digit span: maximum number of digits the subject could recall in reverse order immediately following oral presentation. He or she was allowed two trials at each level and the task was discontinued when both were failed. (6) Paired associate learning: number of trials to reach the criterion of three correct in learning unrelated word pairs. (7) Threshold detection: an array of small rectangles was displayed on a visual display unit. After a brief period of time, an extra rectangle was added. The subject was required to indicate which it was. The ‘threshold’ was the minimum time gap which the subject required to perceive that an extra rectangle has been added. (8) Sedation score: subjects rating of level of alertness using a lo-cm line, where 0 was ‘nearly asleep’ and 10 ‘wide awake’. (9) Side-effect score: the subject was presented with a standardised list of 15 symptoms (e.g., dizziness, diplopia). Each one was graded zero, one, two, three, depending on severity. Individual ratings were summated to produce a total score. TABLE
Analyses Blood samples were taken into heparinised tubes, centrifuged immediately and the plasma stored at -20°C for batch analysis. Total CBZ and VPA concentrations were obtained by enzyme immunoassay (Emit, Syva, Palo Alto, CA, USA) and free CBZ23 and VPAz4 were determined by equilibrium dialysis and enzyme immunoassay. CBZ-E concentrations were assayed by high-performance liquid chromatography as previously describedB. Areas under the concentration-time curve (AUC) for a dosage interval were calculated using the trapezoidal rule. Statistics were performed on antiepileptic drug concentrations using the Wilcoxon signed-rank test for matched pairs. Cognitive function data were examined by repeated-measures two-way analysis of variance, and differences between subgroups by Mann-Whitney U test for unpaired samples. The significance of correlations was explored using linear regression. RESULTS Pharmacokinetics No changes in total or free CBZ concentrations were found following the addition of VPA (Table II). However, CBZ-E concentrations were significantly increased for the whole patient group. This is illustrated for the 17 patients taking CBZ twice daily in Fig. 1. Overall, there was a median 25% increase in mean CBZ-E levels following VPA treatment. However, there was considerable inter-individual variation in the extent of this interaction, ranging from a 25% decrease to an in-
II
Median corb~~epine (‘CBZ) and carob~zep~~e l&U epoxide (CBZ-E) concen~~‘o~ in 24 ep~epticpact addition of sodium valproate 500 mg twice daily for 5 days Total CBZ
Free CBZ
Without VPA
With VPA
95% CI Without for di~ere~ce VPA
CBZ-E With VPA
95% CI for d#erence
Without VPA
With VPA
95% CI for di~~ence
2.1** 2.7.
+0.2 to +0.9
1.5**
+0.1 to +os
20***
+2
Mean (m&I)
8.4
8
-0.3 to +1.2
2
1.8
-0.1 to +0*3
1.3
Maximum (men)
9.4
8.8
-0.1 to +1.5
2.6
2.2
-0.1 to +0.4
1.8
Minimum (mg/l) AUC (me)
6.9 80
6.8 88
-0.6 to +0.9 -4.2 to + 12
1.6 22
1.6 21
-0.2 to +0.2 -0.5 to +4.0
1.1 15
*P<005~**P<001~‘**P<0.001. . f . t
before and after the
+0.1 to +1.0 to +9.5
196 Clrblmarrpina 10
10 -
11 epoxida a
conc~ntrrtionhg/ll
,.Ol 1
a-
T
Number Patients
of 6-
2.6 42.2 21.6 O1.4
1.0 J, 0
2
4
6
6
Hours after crrbamazapine
I 12
10
dose
-25 - 0
0 - 24
25.49
50-74
>75
Fig. 2. Percentage change in median carbamazepine 10,ll epoxide concentrations following the addition of sodium valproate 500 mg twice daily for 5 days in 24 epileptic patients receiving carbamazepine as monotherapy.
Fig. 1. Carbamazepine 10,ll epoxide concentrations over a dosage interval before and after the addition of sodium valproate 500 mg twice daily for 5 days in 17 epileptic patients taking carbamazepine in twice daily dosage.
Cognitive function
crease of 123% in median CBZ-E concentrations (Fig. 2). Mean VPA concentrations over a dosage interval varied widely between patients with a median total VPA of 44 mg/l (range 25-69 mg/l) and a free value of 4.3 mg/l (range 1.6-6.8 mg/l). Mean CBZ-E concentrations in individual patients correlated significantly with mean total (r = 0.5, P < 0.05, 95% CI 0 to +0.08) and free (r = 0.7, P < 0.001,95% CI +0.09 to +0.25) VPA levels.
The results of cognitive function testing in the 16 patients are shown in Table III. The addition of VPA was not associated with any deterioration in performance in any of the tests. However, mean threshold detection scores in the subgroup of patients (n = 7) showing a greater than 25% increase in median CBZ-E concentrations following the addition of VPA were higher than in those (n = 9) with little change (4.8 versus 1.7, P < 0.05). There were no significant differences in performance between these groups for the other tests in the battery (data not shown).
TABLE III Cognitive function tests before and I and 8 h afier the morning carbamazepine
dose with and without additional
sodium
valproate
I6 epileptic patients Test
Decision time (ms)* Movement time (ms)* Finger tapping (per 15 s) Forward digit span Backward digit span Verbal learninga Threshold detectiona (frame units) Sedation scorea (cm) Side-effect score=
Before valproate
After valproate Oh
Ih
8h
(0.3) (0.2) (18) (1.5) (1.6) (4)
0.60 (0.4) 0.31 (0.1) 76 (13) 6.2 (1.5) 4.1 (1.4) 10.8 (5)
0.56 (0.2)
0.57 (0.2) 0.34 (0.1) 75 (17) 5.9 (1.4) 4.7 (1.7) 10.1 (5)
3.2 (3.5) 2.9 (3.0) 9.4 (8)
3.1 (3.0) 2.0 (2.3) 7.2 (7)
3.2 (3.4) 1.8 (2.6) 8.5 (11)
Oh
lh
8h
0.52 (0.2) 0.31(0.1) 74 (20) 6.3 (1.3) 4.1 (1.2) 11.1(5)
0.54 (0.3) 0.32 (0.1) 71 (19) 6.8 (1.1) 4.3 (1.9) 10.2 (4)
0.62 0.34 73 6.3 4.3 9.9
3.3 (3.9) 3.3 (3.0) 13.2 (8)
3 (3.0) 2.1 (2.3) 9.8 (9)
All values are mean (SD). a Denotes test where low score is better than high score.
0.32 (0.1)
71 (17) 5.9 (1.8) 4.5 (1.6) 10.1 (6)
3.3
(3.6)
2.9 7.5
(2.9) (8)
in
197
DISCUSSION Reports of the effect of VPA on total CBZ concentrations are conflicting with a risez6,decrease” or no change%,” noted in various studies. Levy and his colleagues21 reported VPA-mediated inhibition of CBZ metabolism during intravenous infusions in rhesus monkeys. This lack of consensus may be partly explained by a small additional effect of VPA in displacing CBZ from binding sites on plasma proteins20*21,27, Higher free CBZ levels with prolongation of the elimination half-life have been observed in healthy volunteers taking both drugs ” . In the present study, however, in 24 epileptic patients we observed no significant alterations in total or free CBZ concentrations after 5 days’ treatment with a sizeable dose of VPA. The interaction between CBZ and VPA has generated some interest since increased levels of CBZ-E were first recorded in epileptic patients taking both drugs’5,‘6.The effect of VPA on CBZE metabolism is a consequence of inhibition of epoxide hydrolase, the enzyme responsible for its breakdown to the inert dihydrodio117-‘9.Although side-effects experienced by some patients on the combination of CBZ and VPA have been attributed to unsuspected high concentrations of CBZE ‘*,14,other workers have found no relationship between CBZ-E levels and toxic events28. In the only placebo-controlled evaluation, a mean 52% increase in CBZ-E levels was associated with some deterioration in psychomotor performance in healthy volunteers*‘. Considerable variation in the extent of this interaction has been demonstrated in small numbers of patients with epilepsy using different VPA doses. Pisani and colleagues= reported an increase in CBZ-E concentrations between 29% and 23g%, whereas Levy and coauthors*l observed a range of changes from a 14.7% decrease to an 148.4% increase. In our study most patients showed an elevation in CBZ-E concentrations during VPA co-administration. The median increase was 25%. However, there was considerable variation in the magnitude of this
effect, which varied from a 25% decrease to an increase of 123% in mean CBZ-E levels. Part of this va~abi~ty can be accounted for by the substantial differences in circulating VPA concentrations among patients taking the same dose. This may be, in part, due to a combination of concentrationdependent protein binding9 and differences in VPA clearance as a consequence of enzyme induction by CBZ30,3’.The observed correlation between total and free VPA and CBZ-E concentrations suggests that the extent of this effect depends on the VPA dosage. Accordingly, the higher the dose of both drugs, the more likely that this interaction will be clinically relevant. The relationship between seizures, anticonvulsant drugs, psychosocial difficulties and impairment of cognitive function is highly complex. We have shown in previous work that high CBZ-E levels were associated with poorer performance in our battery of cognitive function tests”. In this study we could demonstrate no deterioration in performance of any test after the addition of VPA. Although these observations were uncontrolled and may, therefore, be liable to bias, they do suggest that patients receiving CBZ as monotherapy tolerate the introduction of VPA at this dosage despite an increase, sometimes substantial, in CBZE concentrations. Exploration of the interaction between CBZ and VPA has demonstrated that, although pharmacokinetically significant, there is considerable variation in its magnitude. Although CBZ-E concentrations are not readily predictable in a given patient, it is unlikely that measurement of this metabolite will be helpful in this clinical setting. The adverse effects of this interaction is likely to be small in all but a few patients receiving large doses of both drugs.
Our grateful thanks go to Anne Somers for expert secretarial assistance.
198 REFERENCES 1 Editorial, Carbamazepine update. Lancer, ii (1989) 595-597. 2 Editorial, Sodium valproate, Lancer, ii (1988) 1229- 1231. 3 Aicardi, J., Clinical approach to the management of intractable epilepsy, Dev. Med. Child. Neural., 30 (1988) 429-440. 4 Brodie, 5
6 7
8
M.J., Established anticonvulsants and the treatment of refractory epilepsy, Lancer, 336 (1990) 3.50-354. Shorvon, S.D., Medical assessment and treatment of chronic epilepsy, Br. Med. J., 302 (1991) 363-366. Scheuer, M.L. and Pedley, R.A., The evaluation and treatment of seizures, N. Engf. J. Med., 233 (1991) 1468-1474. Tomson, T., Almkvist, O., Nilsson, B.Y., Svensson, J.O. and Bertilsson L., Carbamazepine IO,11 epoxide in epilepsy. A pilot study, Arch. Neurol., 47 (1990) 888-892. Lindhout, D., Hoppener, R.J.E.A. and Meinardi, H., Teratogenesis of antiepileptic drug combinations with special emphasis on epoxidation (of carbamazepine), Epilepsia, 25 (1984) 77-83.
9 Patsalos, P.N., Stephenson,
T.J., Krishna, S., Elyas, A.A., Lascelles, P.T. and Wiles, C.M., Side-effects induced by carbamazepine 10,ll epoxide, Lancet, ii (1984) 496. 10 Rosenfield, W.E., Holmes, G.B., Graves, N.M. and Peterson, A.L., Carbamazepine 10,lI epoxide serum concentration as a predictor of toxicity, Epilepsia, 28 (1987) 580-581. 11 Gillham, R., Williams, N., Wiedmann, K., Butler, E., Larkin, J.G. and Brodie, M.J., Concentration-effect relationships with carbamazepine and its epoxide on psychomotor and cognitive function in epileptic patients, J. Neurol. Neurosurg. Psychiatry, 51 (1988) 929-933. 12 McKee, P.J. W., Larkin, J.G. and Brodie, M.J., Acute psychosis with carbamazepine and sodium valproate, Lancer, i (1989) 167. 13 Kerr, M. and Levy, R.H., Inhibition of epoxide hydrolase by anticonvulsants and risk of teratogenicity, Lance& i (1989) 610-611. 14 Rambeck, B., Salke-Treumann, A., May, T.H. and Boenigk, H.E., Valproic acid induced carbamazepine 10,ll epoxide toxicity in children and adolescents, Eur. Neurol., 30 (1990) 79-83. 15 McKauge, L., Tyrer, J. and Eadie, M.J., Factors influencing simultaneous concentrations of carbamazepine and its epoxide in plasma, Ther. Drug Monit., 3 (1981) 63-70. 16 Brodie, M.J., Forrest, G. and Rapeport, W.G., Carbamazepine lo,11 epoxide concentrations in epileptics on carbamazepine alone and in combination with other anticonvulsants, Br. .I. Clin. Pharmacol., 16 (1983) 747-750. 17 Chang, S.L. and Levy, R.H., Inhibition of epoxidation of carbamazepine by valproic acid in the isolated perfused rat liver, J. Pharmacokinet. Biopharm., 13 (1985) 453-466. 18 Pisani, F., Caputo, M., Fazio, A., Oteri, G., RUSSO,M., Spina, E., Perucca, E. and Bertilsson, L.. Interaction of
carbamazepine 10,ll epoxide, an active metabohte of carbamazepine, with valproate: a pharmacokinetic study, Epilepsia, 31 (1990) 339-342. 19 Robbins, D.K., Wedlund, P. J., Kuhn, R., Baumann, R.J ,
Levy, R.H. and Chang, S.L., Inhibition of epoxide hydrolase by valproic acid in epileptic patients receiving carbamazepine, Br. J. Clin. Pharmacol., 29 (1990) 759-762, 20 Macphee, G.J.A., Mitchell, J., Wiseman, L., McLellan, A.R., Park, B.K., McInnes, G.T. and Brodie, M.J., Effect of sodium valproate on carbamazepine disposition and psychomotor profile in man, Br. J. Clin. Pharmacol., 25 (1988) 59-66. 21 Levy, R.H., Moreland,
T.A., Morselh, P.L., Guyot, M., Brachet-Liermain, A. and Loiseau, P., Carbamazepinel valproic acid interaction in man and rhesus monkey, Epi-
iepsia, 25 (1984) 338-345. 22 Pisani, F., Fazio, A., Otari, G., Ruello, C., Gitto, F., Rus-
so F. and Perucca, E., Sodium valproate and valpromide: differential interactions with carbamazepine in epileptic patients, Epifepsia, 27 (1986) 548-552. 23 Macphee, G.J.A., Goldie, C., Poulston, D., Potter, L.. Agnew, E., Laidlaw, J. and Brodie, M.J., Effect of carbamazepine on psychomotor performance in naive subjects, Eur. .I. Clin. Pharmacol.,
30 (1986) 37-42.
24 Rapeport,
W.G., Mendelow, A.D., French, G., Macpherson, P., Teasdale, E., Agnew, E., Thompson, G.G. and Brodie, M.J., Plasma protein-binding and CSF concentrations of valproic acid in man following acute oral dosing, Br. J. Clin. Pharmacol.,
16 (1983) 365-369.
25 Macphee,
26 27
28
29
30
G.J.A., Thompson, G.G., Scobie, G.. Agnew. E., Park, B.K., Murray, T., McCall, K.E.L. and Brodie, M.J., Effects of cimetidine on carbamazepine auto- and hetero-induction in man, Br. J. Clin. Pharmacol., 18 (1984) 411-419. Jeavons, P.M. and Clarke, J.E., Sodium valproate in the treatment of epilepsy, Br. Med. J., 2 (1974) 584-586. Mattson, G.F., Mattson, R.H. and Cramer, J.A., Interaction between valproic acid and carbamazepine: an in vitro study of protein binding, Ther. Drug Monir., 4 (1982) 181-184. Theodore, W .H., Narang, P.K., Holmes, M.D., Reeves. P. and Nice, F.J., Carbamazepine and its epoxide: relation of plasma levels to toxicity and seizure control, Ann. Neurol., 25 (1989) 194-196. Bowdle, T.A., Pate& I.H., Levy, R.H., and Wilensky, A.J. Valproic acid dosage and plasma protein binding and clearance, Clin. Pharmacol. Ther., 28 (1981) 485-492. Bowdle, T.A., Levy, R.H. and Cutler, R.E., Effect of carbamazepine on valproic acid kinetics in normal subjects, Clin. Pharmacol. Ther., 26 (1979) 629-634.
31 May T. and Rambeck, B., Serum concentration
acid: influence of dose and co-medication, Monk,
6 (1984) 334-343.
of valproic Ther.
Drug