Valproic Acid in Epilepsy: Clinical and Pharmacological Effects Richard H. Mattson, MD, Joyce A. Cramer,
BS, Peter D. Williamson, MD, and Robert A.
Novelly, PhD
The antiepileptic drug valproic acid was studied in an open clinical trial as adjunct medication for 23 patients with uncontrolled seizures of a g e o e M or partial type. Two-thirds of the patients experienced reduction in seizure frequency ranging from 25 to 100%. Extensive testing revealed no evidence of serious systemic toxicity due to the drug. Minor side effects (e.g., nausea, vomiting, or sedation) were usually transient. Sodium valproate syrup and valproic acid in capsules gave equivalent mean low (23.3 &ml) and maximum (42.5 c18/ml)serum concentrations. The drug had a relatively short half-life of 8.7 hours, necessitating administration in divided daily doses. During initiation of valproate therapy there was evidence of a decline in total serum phenytoin concentration (16.5 to 10.2 pg/ml;p < 0.001)while the percentage of free pbenytoin increased (10.9to 20%). The quantity of unbound phenytoin was relatively stable throughout. This observation was interpreted as a drug interaction: valproic acid competed with phenytoin for access to plasma protein binding sites. Manson RH,Cramer JA, Williamson PD, et al: Valproic acid in epilepsy: clinical and pharmacologicaleffects. Ann Neurol3:20-25.1978
Materials a n d Methods Twenty-three adult patients participated in an open, three-month clinical study; each gave informed consent. The 20 men and 3 women were subject to uncontrolled
seizures despite therapeutic serum concentrations of two or three antiepileptic drugs, and had no evidence of progressive medical or neurological disorder. As documented by case history, clinical observations, and, in many cases, electroencephalogram and videotape recordings, the primary seizure types were: 1 1 partial complex; 5 partial elementary; 2 partial with secondary generalization; 2 generalized tonic-clonic; and 3 generalized absence. Each patient served as his own control based on documentation of seizure frequency for three months prior to the addition of valproate to the medication regimen. In 2 patients absence attacks were quantitated by 24-hour EEG and videotape monitoring before and during valproate therapy. All patients were hospitalized for one week during initiation of valproate therapy for observation of clinical efficacy, toxicity, and pharmacokinetics. Laboratory tests administered prior to and during valproate therapy included blood measurements of the following:alkaline phosphatase, serum glutamic-oxaloacetic transaminase, lactic dehydrogenase, blood urea nitrogen, glucose, calcium, phophorus, sodium, potassium, chloride, bilirubin, total protein, albumin, uric acid, creatinine, cholesterol, white blood cell and red blood cell count, differential, hemoglobin, hematocrit, platelet and reticulocyte counts, and bleeding time. Urinalysis plus 24-hour urine creatinine analysis were done. Each patient also received audiological, ophthalmologic, electrocardiographic, and electroencephalographic examinations and a battery of neuropsychological tests [ 161 at baseline and at three- or six-month intervals. After the initial three-month trial, 14 of 17 patients who benefited from valproic acid agreed to participate in a long-term study. The drug, supplied as Depakene by Abbott Laboratories
From the Epilepsy Center, Veterans AdministrationHospital, West Haven* Yale-New Haven Hospital* and the Of Neurology, Yale University School of Medicine, New Haven, CT.
Accepred for publication June 28, 1977. Address reprint requests to Dr Manson, Epilepsy Center, Vererans Administration Hospital, Haven, CT 065 16,
Valproic acid (dipropylacetic acid) is a promising antiepileptic drug only recently introduced into this country for clinical trial. Meunier and associates [ 141 first reported the anticonvulsant properties of the drug in animals in 1963. Clinical trials by Carraz and co-workers [31 in France and later by investigators in other countries suggested both efficacy and freedom from serious side-effects [2, 8, 12, 21, 221. Valproic acid seems to be most effective in the treatment of generalized seizures of absence type but is of some value for all seizure types [91. Valproic acid has a chemical formula unlike that of other antiepileptic drugs. It is a short-chain, branched fatty acid containing no nitrogen (Fig 1). The exact mechanism of action is unknown, but studies with animals suggest that the drug increases brain levels of y-aminobutyric acid with a corresponding inhibition of seizures [ 5 ] . Although the present clinical study was undertaken to obtain further data concerning the efficacy and toxicity of valproate, it was particularly designed to delineate pharmacological features including bioavailability, half-life, and interaction with other antiepileptic drugs.
20 0364-513417810003-0103$01.25@ 1978 by the American Neurological Association
Fig I . Valproic acid.
(North Chicago, IL)was administered as sodium valproate in a syrup form. After we had a year's experience with this formulation, the drug was given to long-term patients in a capsule as valproic acid. Bioavailabiiity and dosage equivalence were studied. The syrup formulation of sodium valproate (300 mg/5 mi dose) was initially administered three times daily to each patient; the dosage was increased to 900 to 2,400 mg per day as tolerared. The capsules contain 250 mg of valproic acid, which is equivalent to 288 mg of sodium salt. The total daily dosage in capsules ranged from 750 to 2,000 mg per day. Blood concentration of valproic acid was analyzed by the method of Kupferberg (personal communication, 1975). Internal-standard cyclohexanecarboxylicacid was added to 0.5 ml of serum or plasma, followed by acidification with 0.2 ml of 10% perchloric acid. After extraction with 5 ml of chloroform,the samplewas concentrated in isoamyl-acetate. A Varian 2100 gas chromatograph with a column of 10% diethylene glycol succinate o n 80-100 Supelcoport was used for analysis. The injector temperature was 175"C, the detector 200°C. and the column oven was isothermal at 150°C. Since serum phenytoin concentrations fell in many patients after initiation of valproate therapy, changes in protein binding were measured. Protein binding of phenytoin was estimated by plasma ultrafiltration (at 34°C) with Centriflo CF-50 membrane cones (Amicon, Lexington, MA); this method for separation of unbound phenytoin from plasma proteins is comparable to equilibrium dialysis 171. The coefficientof variation for ultrafiltration with the cones of 10 aliquots of a plasma sample from a patient receiving phenytoin was 3.8% Phenytoin concentrations were determined by EMIT enzyme immunoassay (SYVA Corp, Palo Alto, CA) and validated by gas chromatography using a modification of the method of Chang and Glazko [4]. Concentration of the phenytoin metabolite 5-(phydroxyphenyl)-5-phenyihydantoin(HPPH) was measured in urine by gas chromatography [4].
Results The number of seizures during the initial three months of valproate therapy was compared with the previous three-month period for the 23 patients. The mean monthly seizure frequency was reduced by more than 75% in 8 patients, by 50 to 74% in 5 patients, and by 25 to 49% in 4 patients. Six patients had minimal or no (0 to 24%) decrease in seizure occurrence. Nine out of 16 patients having partial seizures with elementary or complex symptomatology experienced 50% or more reduction in seizure frequency. For those patients with multiple seizure types, tonic-clonic attacks had been well controlled prior to valproate therapy.
Optimal benefit was observed in those with absence attacks documented by 24-hour videotape and EEG monitoring. Paroxysmal spike-wave discharges decreased 79 and 96%. Side-effects occurred in 7 patients but were not serious. Only 2 patients had to withdraw during the first month of the trial because of side-effects. Both experienced sedation, ataxia, and nausea. Sedation and gastrointestinal upset were noted most commonly at the initiation of therapy and were usually transient. Ataxia, nystagmus, diplopia, and changes in psychological tests for drug toxicity usually cleared immediately when the valproate dosage was reduced. The side-effects appeared to be direct effects of the valproate and were not due to interaction with other antiepileptic drugs since no increase in the serum levels of other medications was observed. Repeated observations for nearly two years in some patients revealed no hematological, hepatotoxic, or nephrotoxic effects of the drug. Electroencephalographic recordings revealed a decrease in interictal abnormalities, primarily in patients with absence attacks. A decrease of several hertz in background alpha rhythm accompanied the transient clinical complaints of hypnotic toxicity. There was loss of alpha background in 1 patient after the administration of 300 mg and 600 rng doses, at which time increasing side-effects were noted. Toxicity abated as soon as the drug was stopped, and the EEG returned to its pre-treatment pattern. Bioavaika bility The valproate sodium (Vp-Na) was given as syrup in maintenance doses of 900 to 2,400 mg per day (range, 8 to 39 mg/kg). In patients doing well, the dose was not further increased, and the maximal tolerable dose was not fully tested. When 1 1 long-term patients were changed to the valproic acid (Vpa) capsule formulation, doses were approximated one-for-one using the following equivalence: 300 mg V p N A = 260 mg V p a Maintenance doses of Vpa ranged from 750 to 2,000 mg per day (10 to 33 mg/kg). The mean Vpa dose of 18.4 mg per kilogram was little changed from the mean V p N a dose of 20.9 mg per kilogram. Absorption of the drug differed between the two formulations (Table 1). The syrup was rapidly absorbed and often gave peak serum levels within 15 minutes of ingestion. Similar peak levels were achieved somewhat more slowly in nonfasting patients, but within 60 minutes. The rapid elevation of serum concentration was followed by a steady decline, resulting in wide fluctuations in valproic acid serum concentrations throughout the day. The use of gelatin capsules resulted in slower absorption of Vpa; it took 1 to 2 hours to reach a maximum. The serum level occasionally reached a plateau 2 to 4 hours before declin-
Mattson et al: Valproic Acid in Epilepsy 2 1
Table 1 . AbJorptiori a d HaU-Lge of Sodium Vatproale and Vulproii Acid
Sodium Valproace Time TIll,XU
T,r21'.c
Range ' N
Syrup 15-60 rnin 8.73 hr 6-10.5 hr 13
601
50-
Valproic Acid Capsules
40-
60-120 min
8.67 hr 6-10.5 hr
6
"Interval between administration of the oral dose and point of maximum serum concentration. hSerum half-life. 'Serum half-life and range show no significant difference between the two formulations.
I I
30 -
VALPROIC ACID ugh1
I I
I I 1
-600mq .--.600mQ
01
20.
I
6mos day 5
I
1 ing. The range of the half-life of valproic acid in blood was found to be 6 to 10.5 hours (average, 8.7 hours) in these patients receiving two or three other antiepileptic drugs. When the first dose of 600 mg Vp-Na, given on the fifth day, was compared with a dose given after six months of valproate therapy, no change in half-life was observed, as shown in Figure 2. The unaltered rate of elimination indicates no selfinduction of metabolism by valproate. T h e variation in valproic acid serum levels was defined for both formulations by obtaining blood samples serially to determine the maximum level achieved by each patient as well as the lowest level prior to the next dose (approximately one half-life). Table 2 summarizes the mean low and high valproic acid serum concentrations with multiple daily dosing. The mean valproic acid serum concentrations for all doses, both syrup and capsule formulation, were 23.3 ? 8.0 pg per milliliter 8 hours after the previous dose, with a maximum of 42.5 k 11.2 pg per milliliter. Drug Interactioti During the first week of valproate therapy, many patients exhibited a significant reduction in the serum phenytoin level. For 2 1 patients, mean serum concentration of phenytoin before valproate was begun was 19.4 p g per milliliter, representing an average of three o r four previous clinic visits each. The mean serum concentration declined to 14.6 pg per milliliter during the fourth to seventh day after valproate was begun in the hospital (p < 0.001). An average of two or three phenytoin serum levels at the lowest point during four months showed a decline from 19.4 to 11.1 pg per milliliter during valproate therapy ( p < 0.001). T h e rapid fall of the phenytoin level in many patients reflected displacement of phenytoin from plasma protein by valproic acid (Table 3). When valproate was stopped, phenytoin returned to approximately 10% free, considered to be within the normal range [ 101. The reciprocal relationship between total plasma phenytoin and the percentage of free pheny2 2 Annals of Neurology Vol 3
No 1 January 1978
lo-
0
I
2
3
4
5
6
hours
Fig 2 . Elimination of valproir ucid from serum. Vafproate u'as started on duy I at 300 mg per dose ( 3 doses per da.y), increusing to 450 nig per dore o n day 3 and to 600 mfi per dose o n day 5 . Serial bloodsamples were obtained ufter the first dose of 600 rng of Vp-Na on day S and ufter six mnnths of therapy. There is n o apparent difference in the rate of elimination of the drug between initiation of therapy or aber long-term therapy.
toin is illustrated in Figure 3. When Patient 6 was started on valproate treatment at a low dosage, the total phenytoin level was 17.5 pg per milliliter with 8% free. After six months of Vp-Na treatment (2,400 mg/day), the patient was rehospitalized and the valproate was stopped. Mean total phenytoin levels were 14.6 f 1.4 p g p e r milliliter (N = 7) with 11.3 f 2.0% free when the patient was not taking valproate and was receiving 400 mg of phenytoin and 160 mg of phenobarbital daily. When valproate was restarted, total phenytoin immediately declined to a mean of 8.5 2 1.0 pg per milliliter (N = 5) with a concurrent and sustained increase of free phenytoin to 18.5 t 2.7% (both,p < 0.001). The absolute concentration of free phenytoin in plasma remained essentially unchanged whether the patient was on or off valproate therapy. The quantity of unbound phenytoin prior to and during initiation of valproate (Fig 3) was 1.42 f 0.3 1 pg per milliliter (N = 6). When the valproate was stopped, free phenytoin was 1.63 k 0.19 p g per milliliter (N = 7). It remained in a similar range, 1.55 2 0.20 pg per milliliter (N = 5), when the administration of valproate was resumed. Studies were done to see if the decrease in total phenytoin was secondary to altered metabolism re-
Table 2 . Coniparison of Mean Serum Concentrations
of
Valproic Acid in Syrup and Capsule Forma
Dose of Sodium Valproate
Dose of
Valproic Acid Capsule (mg)
Syrup (me)
Determination
300
450
600
250
500
All Doses
16.3 2.0 10
22.2 7.0
26.5 8.9 17
21.1 4.7 5
28.5 7.4 6
23.3 8.0 44
33.0
43.9 8.6
30.9
6
48.7 11.0 18
3
47.7 7.2 4
42.5 11.2 42
20.8
24.9
10.9
21.0
19.2
8 hr
pglml Vpa SD N Max pglrnl Vpa SD N
6
4.3 11
Mdkg
11.2
8.7
aSamples were obtained approximately 8 hours after the dose was administered and serially thereafter to determine the maximum level for each patient. Some patients were tested on more than one dose.
lated to increased absolute phenytoin in serum when displaced by valproic acid. While Patient 6 was under observation, 24-hour urine samples were collected for quantification of the principal phenytoin metabolite H P P H (Fig 4). The samples showed normal H P P H output from 400 mg per day of phenytoin both prior to and during valproate therapy (mean, 244.7 19.5 mg per day; N = 7) except when valproate was being started and stopped. Initiation of valproate therapy brought an immediate, transient increase in H P P H production, which was repeated subsequently when valproate was restarted. The mean H P P H output rose to 342.6 t 25.7 mg per day (N = 4) on these two occasions. Each time, urinary H P P H levels returned
*
Table 3. Protein Binding of Phenytoin following Valproate Administration a
Valproare Dose (mg/day) Determination
0
900
1,350
Mean phenytoin level (pg/ml) Total Free
16.5 1.74
13.3
10.2
10.9 1.8
16.1 5.0
20.0 1.7
11
9 < 0.001
1.94
2.08
5% free phenytoin Mean SD N
P
25
< 0.001
L
BS, Peter D. Williamson, MD, and Robert A.
Novelly, PhD
The antiepileptic drug valproic acid was studied in an open clinical trial as adjunct medication for 23 patients with uncontrolled seizures of a g e o e M or partial type. Two-thirds of the patients experienced reduction in seizure frequency ranging from 25 to 100%. Extensive testing revealed no evidence of serious systemic toxicity due to the drug. Minor side effects (e.g., nausea, vomiting, or sedation) were usually transient. Sodium valproate syrup and valproic acid in capsules gave equivalent mean low (23.3 &ml) and maximum (42.5 c18/ml)serum concentrations. The drug had a relatively short half-life of 8.7 hours, necessitating administration in divided daily doses. During initiation of valproate therapy there was evidence of a decline in total serum phenytoin concentration (16.5 to 10.2 pg/ml;p < 0.001)while the percentage of free pbenytoin increased (10.9to 20%). The quantity of unbound phenytoin was relatively stable throughout. This observation was interpreted as a drug interaction: valproic acid competed with phenytoin for access to plasma protein binding sites. Manson RH,Cramer JA, Williamson PD, et al: Valproic acid in epilepsy: clinical and pharmacologicaleffects. Ann Neurol3:20-25.1978
Materials a n d Methods Twenty-three adult patients participated in an open, three-month clinical study; each gave informed consent. The 20 men and 3 women were subject to uncontrolled
seizures despite therapeutic serum concentrations of two or three antiepileptic drugs, and had no evidence of progressive medical or neurological disorder. As documented by case history, clinical observations, and, in many cases, electroencephalogram and videotape recordings, the primary seizure types were: 1 1 partial complex; 5 partial elementary; 2 partial with secondary generalization; 2 generalized tonic-clonic; and 3 generalized absence. Each patient served as his own control based on documentation of seizure frequency for three months prior to the addition of valproate to the medication regimen. In 2 patients absence attacks were quantitated by 24-hour EEG and videotape monitoring before and during valproate therapy. All patients were hospitalized for one week during initiation of valproate therapy for observation of clinical efficacy, toxicity, and pharmacokinetics. Laboratory tests administered prior to and during valproate therapy included blood measurements of the following:alkaline phosphatase, serum glutamic-oxaloacetic transaminase, lactic dehydrogenase, blood urea nitrogen, glucose, calcium, phophorus, sodium, potassium, chloride, bilirubin, total protein, albumin, uric acid, creatinine, cholesterol, white blood cell and red blood cell count, differential, hemoglobin, hematocrit, platelet and reticulocyte counts, and bleeding time. Urinalysis plus 24-hour urine creatinine analysis were done. Each patient also received audiological, ophthalmologic, electrocardiographic, and electroencephalographic examinations and a battery of neuropsychological tests [ 161 at baseline and at three- or six-month intervals. After the initial three-month trial, 14 of 17 patients who benefited from valproic acid agreed to participate in a long-term study. The drug, supplied as Depakene by Abbott Laboratories
From the Epilepsy Center, Veterans AdministrationHospital, West Haven* Yale-New Haven Hospital* and the Of Neurology, Yale University School of Medicine, New Haven, CT.
Accepred for publication June 28, 1977. Address reprint requests to Dr Manson, Epilepsy Center, Vererans Administration Hospital, Haven, CT 065 16,
Valproic acid (dipropylacetic acid) is a promising antiepileptic drug only recently introduced into this country for clinical trial. Meunier and associates [ 141 first reported the anticonvulsant properties of the drug in animals in 1963. Clinical trials by Carraz and co-workers [31 in France and later by investigators in other countries suggested both efficacy and freedom from serious side-effects [2, 8, 12, 21, 221. Valproic acid seems to be most effective in the treatment of generalized seizures of absence type but is of some value for all seizure types [91. Valproic acid has a chemical formula unlike that of other antiepileptic drugs. It is a short-chain, branched fatty acid containing no nitrogen (Fig 1). The exact mechanism of action is unknown, but studies with animals suggest that the drug increases brain levels of y-aminobutyric acid with a corresponding inhibition of seizures [ 5 ] . Although the present clinical study was undertaken to obtain further data concerning the efficacy and toxicity of valproate, it was particularly designed to delineate pharmacological features including bioavailability, half-life, and interaction with other antiepileptic drugs.
20 0364-513417810003-0103$01.25@ 1978 by the American Neurological Association
Fig I . Valproic acid.
(North Chicago, IL)was administered as sodium valproate in a syrup form. After we had a year's experience with this formulation, the drug was given to long-term patients in a capsule as valproic acid. Bioavailabiiity and dosage equivalence were studied. The syrup formulation of sodium valproate (300 mg/5 mi dose) was initially administered three times daily to each patient; the dosage was increased to 900 to 2,400 mg per day as tolerared. The capsules contain 250 mg of valproic acid, which is equivalent to 288 mg of sodium salt. The total daily dosage in capsules ranged from 750 to 2,000 mg per day. Blood concentration of valproic acid was analyzed by the method of Kupferberg (personal communication, 1975). Internal-standard cyclohexanecarboxylicacid was added to 0.5 ml of serum or plasma, followed by acidification with 0.2 ml of 10% perchloric acid. After extraction with 5 ml of chloroform,the samplewas concentrated in isoamyl-acetate. A Varian 2100 gas chromatograph with a column of 10% diethylene glycol succinate o n 80-100 Supelcoport was used for analysis. The injector temperature was 175"C, the detector 200°C. and the column oven was isothermal at 150°C. Since serum phenytoin concentrations fell in many patients after initiation of valproate therapy, changes in protein binding were measured. Protein binding of phenytoin was estimated by plasma ultrafiltration (at 34°C) with Centriflo CF-50 membrane cones (Amicon, Lexington, MA); this method for separation of unbound phenytoin from plasma proteins is comparable to equilibrium dialysis 171. The coefficientof variation for ultrafiltration with the cones of 10 aliquots of a plasma sample from a patient receiving phenytoin was 3.8% Phenytoin concentrations were determined by EMIT enzyme immunoassay (SYVA Corp, Palo Alto, CA) and validated by gas chromatography using a modification of the method of Chang and Glazko [4]. Concentration of the phenytoin metabolite 5-(phydroxyphenyl)-5-phenyihydantoin(HPPH) was measured in urine by gas chromatography [4].
Results The number of seizures during the initial three months of valproate therapy was compared with the previous three-month period for the 23 patients. The mean monthly seizure frequency was reduced by more than 75% in 8 patients, by 50 to 74% in 5 patients, and by 25 to 49% in 4 patients. Six patients had minimal or no (0 to 24%) decrease in seizure occurrence. Nine out of 16 patients having partial seizures with elementary or complex symptomatology experienced 50% or more reduction in seizure frequency. For those patients with multiple seizure types, tonic-clonic attacks had been well controlled prior to valproate therapy.
Optimal benefit was observed in those with absence attacks documented by 24-hour videotape and EEG monitoring. Paroxysmal spike-wave discharges decreased 79 and 96%. Side-effects occurred in 7 patients but were not serious. Only 2 patients had to withdraw during the first month of the trial because of side-effects. Both experienced sedation, ataxia, and nausea. Sedation and gastrointestinal upset were noted most commonly at the initiation of therapy and were usually transient. Ataxia, nystagmus, diplopia, and changes in psychological tests for drug toxicity usually cleared immediately when the valproate dosage was reduced. The side-effects appeared to be direct effects of the valproate and were not due to interaction with other antiepileptic drugs since no increase in the serum levels of other medications was observed. Repeated observations for nearly two years in some patients revealed no hematological, hepatotoxic, or nephrotoxic effects of the drug. Electroencephalographic recordings revealed a decrease in interictal abnormalities, primarily in patients with absence attacks. A decrease of several hertz in background alpha rhythm accompanied the transient clinical complaints of hypnotic toxicity. There was loss of alpha background in 1 patient after the administration of 300 mg and 600 rng doses, at which time increasing side-effects were noted. Toxicity abated as soon as the drug was stopped, and the EEG returned to its pre-treatment pattern. Bioavaika bility The valproate sodium (Vp-Na) was given as syrup in maintenance doses of 900 to 2,400 mg per day (range, 8 to 39 mg/kg). In patients doing well, the dose was not further increased, and the maximal tolerable dose was not fully tested. When 1 1 long-term patients were changed to the valproic acid (Vpa) capsule formulation, doses were approximated one-for-one using the following equivalence: 300 mg V p N A = 260 mg V p a Maintenance doses of Vpa ranged from 750 to 2,000 mg per day (10 to 33 mg/kg). The mean Vpa dose of 18.4 mg per kilogram was little changed from the mean V p N a dose of 20.9 mg per kilogram. Absorption of the drug differed between the two formulations (Table 1). The syrup was rapidly absorbed and often gave peak serum levels within 15 minutes of ingestion. Similar peak levels were achieved somewhat more slowly in nonfasting patients, but within 60 minutes. The rapid elevation of serum concentration was followed by a steady decline, resulting in wide fluctuations in valproic acid serum concentrations throughout the day. The use of gelatin capsules resulted in slower absorption of Vpa; it took 1 to 2 hours to reach a maximum. The serum level occasionally reached a plateau 2 to 4 hours before declin-
Mattson et al: Valproic Acid in Epilepsy 2 1
Table 1 . AbJorptiori a d HaU-Lge of Sodium Vatproale and Vulproii Acid
Sodium Valproace Time TIll,XU
T,r21'.c
Range ' N
Syrup 15-60 rnin 8.73 hr 6-10.5 hr 13
601
50-
Valproic Acid Capsules
40-
60-120 min
8.67 hr 6-10.5 hr
6
"Interval between administration of the oral dose and point of maximum serum concentration. hSerum half-life. 'Serum half-life and range show no significant difference between the two formulations.
I I
30 -
VALPROIC ACID ugh1
I I
I I 1
-600mq .--.600mQ
01
20.
I
6mos day 5
I
1 ing. The range of the half-life of valproic acid in blood was found to be 6 to 10.5 hours (average, 8.7 hours) in these patients receiving two or three other antiepileptic drugs. When the first dose of 600 mg Vp-Na, given on the fifth day, was compared with a dose given after six months of valproate therapy, no change in half-life was observed, as shown in Figure 2. The unaltered rate of elimination indicates no selfinduction of metabolism by valproate. T h e variation in valproic acid serum levels was defined for both formulations by obtaining blood samples serially to determine the maximum level achieved by each patient as well as the lowest level prior to the next dose (approximately one half-life). Table 2 summarizes the mean low and high valproic acid serum concentrations with multiple daily dosing. The mean valproic acid serum concentrations for all doses, both syrup and capsule formulation, were 23.3 ? 8.0 pg per milliliter 8 hours after the previous dose, with a maximum of 42.5 k 11.2 pg per milliliter. Drug Interactioti During the first week of valproate therapy, many patients exhibited a significant reduction in the serum phenytoin level. For 2 1 patients, mean serum concentration of phenytoin before valproate was begun was 19.4 p g per milliliter, representing an average of three o r four previous clinic visits each. The mean serum concentration declined to 14.6 pg per milliliter during the fourth to seventh day after valproate was begun in the hospital (p < 0.001). An average of two or three phenytoin serum levels at the lowest point during four months showed a decline from 19.4 to 11.1 pg per milliliter during valproate therapy ( p < 0.001). T h e rapid fall of the phenytoin level in many patients reflected displacement of phenytoin from plasma protein by valproic acid (Table 3). When valproate was stopped, phenytoin returned to approximately 10% free, considered to be within the normal range [ 101. The reciprocal relationship between total plasma phenytoin and the percentage of free pheny2 2 Annals of Neurology Vol 3
No 1 January 1978
lo-
0
I
2
3
4
5
6
hours
Fig 2 . Elimination of valproir ucid from serum. Vafproate u'as started on duy I at 300 mg per dose ( 3 doses per da.y), increusing to 450 nig per dore o n day 3 and to 600 mfi per dose o n day 5 . Serial bloodsamples were obtained ufter the first dose of 600 rng of Vp-Na on day S and ufter six mnnths of therapy. There is n o apparent difference in the rate of elimination of the drug between initiation of therapy or aber long-term therapy.
toin is illustrated in Figure 3. When Patient 6 was started on valproate treatment at a low dosage, the total phenytoin level was 17.5 pg per milliliter with 8% free. After six months of Vp-Na treatment (2,400 mg/day), the patient was rehospitalized and the valproate was stopped. Mean total phenytoin levels were 14.6 f 1.4 p g p e r milliliter (N = 7) with 11.3 f 2.0% free when the patient was not taking valproate and was receiving 400 mg of phenytoin and 160 mg of phenobarbital daily. When valproate was restarted, total phenytoin immediately declined to a mean of 8.5 2 1.0 pg per milliliter (N = 5) with a concurrent and sustained increase of free phenytoin to 18.5 t 2.7% (both,p < 0.001). The absolute concentration of free phenytoin in plasma remained essentially unchanged whether the patient was on or off valproate therapy. The quantity of unbound phenytoin prior to and during initiation of valproate (Fig 3) was 1.42 f 0.3 1 pg per milliliter (N = 6). When the valproate was stopped, free phenytoin was 1.63 k 0.19 p g per milliliter (N = 7). It remained in a similar range, 1.55 2 0.20 pg per milliliter (N = 5), when the administration of valproate was resumed. Studies were done to see if the decrease in total phenytoin was secondary to altered metabolism re-
Table 2 . Coniparison of Mean Serum Concentrations
of
Valproic Acid in Syrup and Capsule Forma
Dose of Sodium Valproate
Dose of
Valproic Acid Capsule (mg)
Syrup (me)
Determination
300
450
600
250
500
All Doses
16.3 2.0 10
22.2 7.0
26.5 8.9 17
21.1 4.7 5
28.5 7.4 6
23.3 8.0 44
33.0
43.9 8.6
30.9
6
48.7 11.0 18
3
47.7 7.2 4
42.5 11.2 42
20.8
24.9
10.9
21.0
19.2
8 hr
pglml Vpa SD N Max pglrnl Vpa SD N
6
4.3 11
Mdkg
11.2
8.7
aSamples were obtained approximately 8 hours after the dose was administered and serially thereafter to determine the maximum level for each patient. Some patients were tested on more than one dose.
lated to increased absolute phenytoin in serum when displaced by valproic acid. While Patient 6 was under observation, 24-hour urine samples were collected for quantification of the principal phenytoin metabolite H P P H (Fig 4). The samples showed normal H P P H output from 400 mg per day of phenytoin both prior to and during valproate therapy (mean, 244.7 19.5 mg per day; N = 7) except when valproate was being started and stopped. Initiation of valproate therapy brought an immediate, transient increase in H P P H production, which was repeated subsequently when valproate was restarted. The mean H P P H output rose to 342.6 t 25.7 mg per day (N = 4) on these two occasions. Each time, urinary H P P H levels returned
*
Table 3. Protein Binding of Phenytoin following Valproate Administration a
Valproare Dose (mg/day) Determination
0
900
1,350
Mean phenytoin level (pg/ml) Total Free
16.5 1.74
13.3
10.2
10.9 1.8
16.1 5.0
20.0 1.7
11
9 < 0.001
1.94
2.08
5% free phenytoin Mean SD N
P
25
< 0.001
L
