Blaise F. D. Bourgeois MD
Professor of Neurology, Harvard Medical School; and Director, Division of Epilepsy and Clinical Neurophysiology, Children's Hospital Boston, Boston, Massachusetts
Since its first clinical use in France in 1964 (20), valproate, or valproic acid (VPA), rapidly established itself worldwide as a major antiepileptic drug against several types of seizures. It was soon recognized as a highly effective first-line drug against the generalized seizures encountered in idiopathic or primary generalized epilepsies: absence, generalized tonic-clonic, and myoclonic seizures (58). Clinical experience with VPA has continued to grow and is still the subject of numerous studies and publications. Antiepileptic therapy with VPA has been the subject of several reviews (14,42,79,91).
SPECTRUM OF EFFICACY
Evidence from Randomized Controlled Clinical Trials
When VPA was released for clinical use in North America in 1978, the primary indication was the treatment of absence seizures. The efficacy of VPA and ethosuximide in the treatment of absence seizures was found to be equal in at least two comparison studies (19,86). In a double-blind, crossover study of 16 patients not previously treated for absence seizures and in 29 treatment-refractory patients, the measure of efficacy was based on the frequency and duration of 3/sec generalized spike-and-wave bursts on 12-hour telemetered electroencephalographic (EEG) recordings (86). In the patients who had not previously received antiepileptic therapy, VPA and ethosuximide were equally effective.
The most comprehensive assessment of VPA in the treatment of partial and secondarily generalized seizures was carried out by Mattson et al. (70), in what can be qualified as a landmark study. In a multicenter, double-blind trial, 480 adults with complex partial or secondarily generalized tonic-clonic seizures were randomly assigned to monotherapy with carbamazepine or VPA. Targeted serum levels were 7 to 8 µg/mL for carbamazepine and 80 to 100 µg/mL for VPA. The following efficacy indicators were determined at 12 and 24 months of treatment: seizure count per 12 months, seizure rate per month, percentage of patients without seizure, seizure rating score, the time to first seizure, and a seizure rating score representing a weighted sum of generalized tonic-clonic, complex partial, and simple partial seizures. Systemic toxicity and neurotoxicity were also quantified, and they were combined with the efficacy variables into a composite score (0 to 20 representing a good clinical outcome, and >50 representing an unacceptable outcome). No difference in efficacy between carbamazepine and VPA was found for secondarily generalized tonic-clonic seizures. In the treatment of complex partial seizures, four of five efficacy indicators were significantly in favor of carbamazepine. The composite score for efficacy and toxicity was the same for both drugs in patients with generalized tonic-clonic seizures. In the group of patients with complex partial seizures, the composite score was significantly better for carbamazepine (6.8) than for VPA (16.0) at 12 months, but not at 24 months. From these results, it appears that VPA is one of the drugs of choice in the treatment of secondarily generalized tonic-clonic seizures, and it is a valuable alternative in the treatment of complex partial seizures. Although the composite score is an excellent indicator of efficacy for the comparison of drugs, it is difficult to attribute a relative weight to different side effects, and this relative weight can vary from one patient to another.
These additional randomized studies confirmed that VPA is a valuable first-line treatment for newly diagnosed partial epilepsy in adults. Callaghan et al. (18) compared monotherapy with carbamazepine, phenytoin, and VPA in 181 previously untreated patients, with a median follow-up
ranging from 14 to 24 months. In the 79 patients with simple or complex partial seizures, there was no significant difference among the three drugs with regard to both seizure reduction and complete control of seizures. In another series of 140 adults with previously untreated seizures, monotherapy with phenytoin and VPA was compared in a randomized design; 76 patients had tonic-clonic seizures only, and 64 patients had predominantly complex partial seizures (99). Using time to 2-year remission and time to first seizure as the measures of efficacy, the authors could find no difference between the two drugs in either group. Among patients with partial seizures, 27% had a 2-year remission while they were receiving VPA, and 29% had a similar remission while they were taking phenytoin. In the trial by Richens et al. (81a), conducted in 300 patients with primarily generalized tonic-clonic seizures or partial seizures, with or without secondary generalization, VPA was equally effective as carbamazepine, regardless of seizure type; however, carbamazepine was associated with a higher withdrawal rate because of adverse events (15% versus 5% on VPA), and therefore more patients taking VPA than carbamazepine (90% versus 75%) remained on the allocated treatment for at least 6 months.
Two comparative studies of VPA were carried out in children (30,103); 260 children with newly diagnosed primary generalized or partial epilepsy were randomized to VPA or carbamazepine and were followed-up for 3 years (103). The doses were titrated as needed and as tolerated according to clinical response. Equal efficacy was found for the two drugs against generalized and partial seizures, and adverse events were mostly mild for both drugs. Four drugs—phenobarbital, phenytoin, carbamazepine, and VPA—were compared in 167 children with untreated tonic-clonic or partial seizures entered into a randomized, unblinded study (30). Based on time to first seizure and to 1-year remission, there was no difference in efficacy at 1, 2, or 3 years. Unacceptable side effects necessitating withdrawal occurred in six of 10 patients receiving phenobarbital, which was prematurely eliminated from the study; such side effects occurred in 9% of children receiving phenytoin and in 4% each of children taking carbamazepine or VPA.
More recently, VPA was evaluated in 143 adult patients with poorly controlled partial epilepsy who were randomized to monotherapy with VPA at low levels (25 to 50 mg/L) or high levels (80 to 150 mg/L) (8). The reduction in the frequency of both complex partial and secondarily generalized tonic-clonic seizures was significantly higher among patients in the high-level group.
In a study of patients with complex partial seizures not controlled by carbamazepine or phenytoin, patients were randomized to add-on VPA or add-on placebo (110). In the intent-to-treat analysis of 137 patients, those treated with VPA experienced a median reduction of 7.9 complex partial seizures per 8 weeks, and 38% of them had a seizure reduction of ≥50%. The corresponding numbers for the placebo group were 2.5 and 19%, respectively, both differences being statistically significant.
Evidence from Other Studies
A reduction of spike-and-wave discharges was demonstrated repeatedly when VPA was administered to patients with typical and atypical absences (1,6,11,67,71). In a group of 25 patients with absence seizures who were treated with VPA for 10 weeks, 19 patients experienced a reduction in seizure frequency, and 21 had a reduction in the total time of spike-and-wave discharge (105). The reduction was >75% in 11 patients. In various series, complete control of simple absence seizures was achieved with VPA monotherapy in 10 of 12 patients (46), 11 of 12 patients (26), 14 of 17 patients (37), and 20 of 21 patients (10). Complete control of absence seizures appears to be more likely when they occur alone than when they occur in combination with another seizure type (10,46). When atypical or “complex” absences are included, results are generally less favorable than in patients with simple absences only (26,35). Absence seizures refractory to VPA and ethosuximide given alone may respond well to a combination of these drugs (85a). VPA and lamotrigine in combination have also been reported to be valuable in refractory absence seizures (36a). Oral VPA was also found to be highly effective in the prevention of recurrent absence status epilepticus (7). In an open evaluation of 25 patients, the yearly frequency of attacks of absence status was decreased from an average of 5.7 to 0.6. The results were clearly better in the 18 patients who had primary generalized epilepsy. Experience has also suggested that VPA is a useful drug in the treatment of myoclonic absence epilepsy (104).
Generalized Tonic-Clonic Seizures
Although it was initially used for the treatment of absences, VPA has since been recognized to be effective in convulsive seizures as well (34, 80, 93, 100). Therapy with VPA as the only drug was used in 36 patients with primary tonic-clonic seizures, of whom 24 had been previously treated with other antiepileptic drugs (26). Complete seizure control was achieved in 33 patients. In a series of 100 children with intractable epilepsy, grand mal seizures were completely controlled by the addition of VPA in 14 of 42 patients (46). Wilder et al. (109) compared VPA with phenytoin in 61 previously untreated patients with generalized tonic-clonic, clonic, or tonic seizures in a randomized fashion. No seizures recurred in 73% of patients treated with VPA and in 47% of patients treated with phenytoin. These percentages increased to 82% for VPA and 76% for phenytoin when seizures that had occurred at a time when therapeutic plasma drug levels had not yet been reached were discounted. In another randomized study of the relative efficacy
of VPA and phenytoin, a 2-year remission was used as the end point (99). In patients with previously untreated tonic-clonic seizures, this remission was achieved in 27 of 37 cases with VPA and in 22 of 39 cases with phenytoin. In the 3-year randomized trial carried out by Richens et al. (81a), patients with primarily generalized tonic-clonic seizures achieved higher 12-month remission rates with VPA than with carbamazepine (76% versus 62%). The average dosage of VPA required to control primarily generalized tonic-clonic seizures was also lower than that required to control partial seizures (821 versus 1,066 mg/day). In two studies of monotherapy for primary (or idiopathic) generalized epilepsies with VPA, complete control of generalized tonic-clonic seizures was achieved in 51 of 70 patients (37) and in 39 of 44 patients (10) in whom only these types of seizures occurred. Excellent results for generalized tonic-clonic seizures were also obtained with VPA monotherapy in children (33).
VPA is the drug of first choice against myoclonic seizures occurring in patients with primary or idiopathic generalized epilepsies (10,26,37). Sixteen of 23 patients with myoclonic epilepsy of adolescence had complete seizure control with VPA monotherapy (26). Among 22 patients with myoclonic epilepsy of adolescence and abnormality on intermittent photic stimulation, 17 had complete seizure control, although 17 patients had not responded to previous medications (26). In general, photosensitivity is easily controlled by VPA, whether it is associated with tonic-clonic, absence, or myoclonic seizures (57). A good response to VPA in the treatment of juvenile myoclonic epilepsy was reported by Delgado-Escueta and Enrile-Bacsal (29). In a study of VPA monotherapy for primary generalized epilepsies, myoclonic seizures were suppressed in 18 of 22 patients (10). Twenty of these 22 patients had another type of seizure (either absence or tonic-clonic) in addition to the myoclonic seizures. Good results with VPA monotherapy were also obtained in children with the so-called benign myoclonic epilepsy of infancy (33). VPA has also been used successfully against the postanoxic intention myoclonus (15,83) and, in conjunction with clonazepam, against the myoclonic and tonic-clonic seizures occurring in patients with severe progressive myoclonic epilepsy (51).
Lennox-Gastaut Syndrome and Infantile Spasms
Like all other antiepileptic therapies that have been used so far, therapy with VPA has been less successful for the generalized seizures encountered in the symptomatic generalized epilepsies (e.g., Lennox-Gastaut syndrome) and for the prevention of infantile spasms occurring as part of West's syndrome. There is also much less information available on the use of VPA in these forms of epilepsy than in the treatment of the primary generalized epilepsies. In the series by Covanis et al. (26), myoclonic absence seizures were fully controlled in three of six patients, but none of the patients was treated with VPA alone. Among their 38 patients with myoclonic astatic epilepsy (a term used by the authors synonymously with Lennox syndrome), only seven patients became and remained seizure free with VPA. However, the addition of VPA was associated with a 50% to 80% improvement in one-third of the patients, and other antiepileptic drugs could be withdrawn or reduced after the introduction of VPA. In their series of 100 children treated with VPA, Henriksen and Johannessen (46) reported that 12 of 27 children with “absences and other seizures” and nine of 39 children with atonic seizures became seizure free.
Several studies on the use of VPA for the prevention of infantile spasms were based either on small numbers of patients (5,76,82) or on a combination of VPA and corticotropin given simultaneously (12,112). In a series of 19 babies with infantile spasms (3), VPA was not used simultaneously with corticotropin. Good spasm control was achieved with VPA as a first drug in eight patients, and these patients therefore did not require corticotropin. The doses of VPA ranged from 20 to 60 mg/kg/day. An analysis of the cases of initial failure with either corticotropin or VPA and subsequent treatment with the other drug revealed a tendency toward a better response to corticotropin. However, treatment with VPA was associated with a lower incidence and severity of side effects. In another study, VPA was given at a dosage of 20 mg/kg/day to 18 infants with infantile spasms who were never treated with corticotropin (78). Short-term results were good to excellent in 12 patients. Follow-up revealed that seven patients still had residual seizures, and 16 had either moderate or severe mental retardation. The results were judged to be similar to those obtained with corticotropin or steroids, but VPA was found to have fewer side effects.
Additional information on the effect of VPA against partial seizures is derived in part from small numbers of patients in studies not dealing primarily with partial seizures. In the series by Covanis et al. (26), nine patients with simple partial seizures responded poorly to VPA, and five of 11 patients with complex partial seizures were seizure free while they were taking VPA, mostly in combination with carbamazepine. Among the 100 children with uncontrolled epilepsy reported by Henriksen and Johannessen (46), 13 had simple partial seizures. Of those, only one became seizure free after the addition of VPA, and a seizure reduction of >75% and 50% to 75% was observed in three patients each. Among 19 patients with complex partial seizures, four became seizure free, five experienced a seizure reduction of 75%, and five experienced a seizure reduction of 50% to
75%. In 24 adults with poorly controlled complex partial seizures, Bruni and Albright (13) added VPA to the existing regimen. A >50% seizure reduction was initially achieved in 12 patients, but in seven patients this improvement was temporary. Thus, a long-term benefit was obtained in only five of 24 patients. Better results were observed when VPA was compared with carbamazepine in an open study of 31 previously untreated patients with partial seizures (65). Seizures were completely controlled by VPA in 11 patients and by carbamazepine in eight patients. However, only 19 patients were followed for a period of 1 year.
Gupta and Jeavons (44) compared 40 patients with complex partial seizures who responded to carbamazepine monotherapy with 45 patients who were not seizure free while receiving carbamazepine monotherapy and in whom VPA was added as a second drug. Their response to therapy was analyzed as a function of the side of the interictal EEG abnormality. Carbamazepine alone fully controlled the seizures in 24 of 47 patients who had a left temporal EEG abnormality exclusively and in only seven of 33 patients with a right-sided abnormality. Among those patients whose seizures were not controlled by carbamazepine alone, the addition of VPA was associated with full seizure control in 18 of the 26 patients with a right temporal abnormality and in only one of the 13 patients with a left temporal abnormality. Although these results suggest a better response to carbamazepine alone in patients with a left temporal abnormality and a better response to the addition of VPA to carbamazepine in patients with a right temporal abnormality, a control group with VPA as the only drug was not included in the study. VPA monotherapy in partial seizures was evaluated retrospectively in 30 patients with simple and complex partial seizures who had not tolerated or had not responded to conventional initial drugs (27). Change to VPA monotherapy was associated with surprisingly good results. Twelve patients became seizure free, 10 experienced reduced seizure frequency by more than 50%, and nine were not improved. All generalized components of partial seizures were controlled in these patients. In patients with refractory partial seizures, as in other types of refractory seizures, the combination of VPA and lamotrigine has been found to produce more favorable results than other antiepileptic drug combinations (79a).
The effectiveness of VPA in the prevention of febrile convulsions is well documented (21,75). In a group of 196 children with febrile seizures of whom 69 received prolonged treatment with phenobarbital, 32 with primidone, and 95 with VPA, Minagawa and Miura (72) found no statistical difference in the recurrence rate in the three groups. In other series of children with febrile seizures, VPA was found to be more effective than phenobarbital, placebo, or no treatment (47,60,68).
Although VPA is effective in the prophylaxis of febrile seizures, its use cannot be recommended for this indication. The use of long-term prophylactic treatment in patients with febrile seizures has declined markedly in recent years. This has been mainly the result of a reassessment of the risk:benefit ratio and of the increasing use of intermittent diazepam during febrile episodes. The latter approach was found to be as effective as prophylactic treatment with VPA in a group of children with a high risk of recurrence of febrile seizures (61).
VPA has also been used for the treatment of seizures during the neonatal period. Two newborns whose seizures were not controlled by conventional antiepileptic drugs responded well to a rectal infusion of VPA (95). In a more recent trial, VPA was administered orally as syrup to six neonates who had persistent seizures despite levels of phenobarbital of >40 µg/mL and despite the administration of an additional anticonvulsant in five of the six infants (38). The loading dose of VPA was 20 to 25 mg/kg, and the initial maintenance dose was 5 to 10 mg/kg every 12 hours. The seizures were controlled in all but one infant, a 30-week gestational age newborn with meningitis. An elevation of serum ammonia was observed in all patients. This elevation was reversible despite continuation of the VPA therapy in three cases. Seizures recurred in two cases after VPA was discontinued because of hyperammonemia. The mean VPA elimination half-life in five of these newborns was 26.4 hours, a value considerably higher than in children and adults.
Since it became available for intravenous administration, VPA has been administered for the treatment of status epilepticus. Among 23 patients treated with an intravenous bolus of 15 mg/kg followed by 1 mg/kg/hr, status epilepticus resolved in 19 (39). Eight patients had tonic-clonic status, and the remainder had absence, myoclonic, focal motor, and focal myoclonic status. Intravenous VPA was also used successfully in two patients with juvenile myoclonic epilepsy who were in myoclonic status epilepticus (90). In a group of 41 children with various forms of status epilepticus, intravenous VPA was successful in 78% overall (101). None of the patients with epilepsia partialis continua responded, but the success rate was 67% to 90% in the other groups.
MODE OF USE
Indications and Current Role in Epilepsy Management
VPA is the prime example of a broad-spectrum antiepileptic drug. It has been shown to be potentially effective against every known type of seizure, including neonatal seizures, febrile seizures, and status epilepticus. VPA is clearly the drug of first choice in idiopathic (primary) generalized epilepsies, including juvenile myoclonic epilepsy, juvenile absence epilepsy, childhood absence epilepsy,
myoclonic absence epilepsy, and myoclonic astatic epilepsy. It is also often used as a drug of first choice in cryptogenic or symptomatic generalized epilepsies such as Lennox-Gastaut syndrome, as well as in most cases of myoclonic seizures, regardless of the syndrome. In patients with infantile spasms, VPA is often used as a drug of second choice, occasionally as a first drug. VPA is used as a drug of first or second choice against generalized tonic-clonic seizures without evidence of focal onset. Against simple or complex partial seizures without or with secondary generalization, VPA is commonly used as a drug of second choice, in certain countries as a first drug. VPA is often used in patients with benign focal epilepsies of childhood. It is currently uncommon for VPA to be used in neonatal seizures. Risk:benefit considerations preclude VPA from being used in the prophylaxis of febrile seizures.
In various parts of the world, the drug is marketed as VPA, sodium VPA, magnesium VPA, sodium hydrogen divalproate, or valpromide, the amide of VPA. VPA is available as oral syrup, immediate-release formulations, and enteric-coated tablets or sprinkles. Slow-release oral preparations and intravenous formulations are available in certain countries. Although VPA syrup has been used for rectal administration (24,98), sodium VPA suppositories are available in certain countries. In volunteers, the relative bioavailability of VPA suppositories was 80% compared with oral syrup, and the time to maximal concentration was longer (3.1 hours versus 1.0 hour) (50). In another study, suppositories were found to be well tolerated, even when administered for several days, and to have the same bioavailability as the oral preparations (56). Comparison of enteric-coated sprinkles with syrup in 12 children showed no difference in overall bioavailability, but the absorption of VPA was slower from sprinkles. The time to maximal concentration was 4.2 hours with sprinkles and 0.9 hour with syrup (23). Divalproex sodium extended-release tablets have become available, which are intended for once-a-day oral administration. Compared with delayed-release tablets given twice daily, these tablets were found to have a relative bioavailability of 81% to 89%, and they produced fluctuations of serum levels that were 10% to 20% lower (unpublished data, Abbott Laboratories, North Chicago, IL).
The question of the appropriate daily dose of VPA has been addressed in numerous reports. The recommended initial dose is 10 to 15 mg/kg/day. The dose may be increased as necessary, and as tolerated, by 5 to 10 mg/kg/day at weekly intervals. Because of pronounced pharmacokinetic interactions, the dose of VPA will differ markedly between patients receiving VPA monotherapy and patients receiving combination therapy, if similar concentrations are to be achieved. Furthermore, the adequate VPA dose or concentration can be a function of the patients seizure type (66,81a). In patients receiving VPA monotherapy, doses between 10 and 20 mg/kg/day usually achieve a good clinical response and result in concentrations within the generally accepted therapeutic range (10,37, 46,109). Because of age-dependent kinetics, younger children may require higher doses (26,81). Patients receiving combination therapy will almost invariably need higher doses if the same concentrations are to be obtained, usually between 30 and 60 mg/kg/day. This effect of comedication is particularly pronounced in children, in whom VPA concentrations in the accepted therapeutic range often cannot be achieved, even with doses >100 mg/kg/day (46).
If the therapeutic levels of VPA are to be achieved rapidly or in patients who are unable to take oral medications, VPA can be administered intravenously (31). This route has also been suggested for the treatment of status epilepticus, with a loading dose of 15 mg/kg (given at a rate of 20 mg/min) followed by 1 mg/kg/hr (39). An alternative loading dose of 20 mg/kg a rate of 33.3 to 555 mg/min (62), or ≤6 mg/kg/min (108), has been advocated. Overall, such rapid intravenous loading with VPA has been well tolerated (74,92). In those receiving intravenous replacement therapy or bolus administration, subsequent doses should be given within 6 hours because of the precipitous fall in VPA levels.
A curvilinear relationship between VPA dose and concentration was reported by Gram et al. (43), with relatively smaller increases in concentrations at higher doses. Among the different possible explanations, the most likely seems to be a higher free fraction at higher concentrations, resulting in a higher total drug clearance. This explanation implies that the relationship between dosage and concentration of free, pharmacologically active drug is in fact linear. When therapeutic levels of VPA are to be achieved rapidly, a loading dose of 12.5 mg/kg given orally has been shown to be adequate (84). Although it is a common practice to divide the daily dose of VPA into two or three single doses because of the short elimination half-life, equally good results were obtained when VPA was administered as a single daily dose (26,40,94). In patients with primary generalized epilepsy, a mean VPA dose of 15.6 mg/kg/day administered once daily was adequate (94). Even though maximal concentrations were about twice as high as minimal concentrations, side effects were rare. Sixteen patients with juvenile myoclonic epilepsy were randomized in a double-blind fashion to monotherapy with VPA at 1,000 and 2,000 mg/day (96,97). The higher dose, with a mean level of 700 µmol/L, did not result in better seizure control than 1,000 mg/day with a mean level of 470.4 µmol/L.
What is the concentration-effect relationship for VPA, and what is the value of determining plasma or serum concentrations? There are two main reasons for the difficulties associated with interpretation of VPA levels obtained in patients: first, VPA levels fluctuate considerably during a 24-hour period because of the short half-life; and second, no good correlation between VPA levels and clinical effects
at a given time has been demonstrated so far. Studying the photoconvulsive response in humans, Rowan et al. (85) found that reduction or abolition of photosensitivity in the EEG occurred on the average 3 hours after peak serum VPA levels had been achieved and persisted for hours as the levels continued to decrease. Burr et al. (17) measured the course of spike-and-wave discharge rate in the EEG and analyzed its relationship with fluctuations in VPA levels. They found no significant correlation, whether they analyzed the actual discharge profile or the profile of changes in discharge rate. During long-term VPA therapy, a continued reduction of spike-and-wave activity or seizure frequency occurred even after a steady state had been reached (10, 12,14). This phenomenon is probably not the result of delayed brain penetration, selective concentration, or slow elimination of VPA from the brain, because a good correlation between the brain and cerebrospinal fluid concentration and free plasma concentrations could be demonstrated in patients undergoing neurosurgery (102). In a more recent report, VPA was found to have the lowest brain: blood ratio of all major antiepileptic drugs, with a fourfold range among patients (89).
In a group of 25 patients with absence seizures, Villareal et al. (105) found no correlation between plasma VPA concentrations and EEG changes, but these investigators observed a clinical response when levels reached 50 to 60 µg/mL (347 to 417 µmol/L). Among children with absence seizures, those who became seizure free during VPA monotherapy had a mean plasma concentration of 83.1 µg/mL (VPA as a first drug) and 59.6 µg/mL (patients previously refractory to ethosuximide) (86). In a group of 28 children whose seizures were completely controlled by VPA monotherapy, the mean plasma level was 65.1 µg/mL (452 µmol/L) (59). Farrell et al. (36) found serum levels of 140 to 420 µmol/L (20.2 to 50.5 µg/mL) by gas chromatography and 210 to 560 µmol/L (30.2 to 80.6 µg/mL) by enzyme-multiplied immunoassay test (EMIT) in 80% of children with complete seizure control. Ishikawa et al. (52) reported mean levels of 47.8 to 85.2 µg/mL (322 to 592 µmol/L) in children whose seizures were well controlled. Younger children had lower average concentrations, and half of them had their seizures controlled with levels of <50 µg/mL (350 µmol/L). In the series by Covanis et al. (26), mean serum levels in seizure-free patients ranged from 82 to 109.5 µg/mL (569 to 760 µmol/L), and these authors recommend a therapeutic range of 60 to 120 µg/mL (417 to 833 µmol/L). In seven children with absence epilepsy, VPA concentrations of 440 to 660 µmol/L (63 to 95 µg/mL) were necessary to achieve a reduction of EEG seizure discharges by ≥50% (11). There is no evidence in any of the studies cited earlier that VPA was actually increased to the minimally effective dose or concentration. In the series by Henriksen and Johannessen (46), clinical effect was observed only after the fasting serum VPA levels had reached values of about 300 µmol/L (43.2 µg/mL), and seizure control was frequently achieved only after these levels had persisted for 2 to 4 weeks. Lethargy and drowsiness were noticed at levels >600 µmol/L in almost all cases. In previously untreated adults, adverse effects were common when plasma VPA levels were >100 µg/mL (700 µmol/L) during monotherapy, and primarily generalized tonic-clonic seizures did not occur at levels of >50 µg/mL (350 µmol/L) (98). A lower limit for the prevention of partial seizures could not be determined in this study.
Based on their finding of a good linear correlation between VPA dose and level in patients receiving monotherapy, Lundberg et al. (66) considered that monitoring VPA levels was not necessary in these patients. In a group of 88 children receiving VPA monotherapy, those with certain side effects had received significantly higher doses than patients without side effects (48). However, a relationship between the incidence of side effects and plasma VPA levels could not be established, except for children presenting with lassitude and drowsiness, whose levels were significantly higher (mean values of 94.5 and 80.4 µg/mL, respectively). In their study on the prevention of febrile seizures, Herranz et al. (47) found no difference in doses or plasma levels among patients with or without recurrence of seizures or side effects. However, patients with side effects were receiving a higher VPA dose. Schobben et al. (87) also found that the therapeutic effect correlated better with the VPA dose in milligrams per kilogram of body weight than with the plasma concentration. Gram et al. (43) concluded that seizure control was better at serum levels of 300 to 350 µmol/L (43 to 50 µg/mL) than at lower values, but they found no correlation between VPA levels and side effects. The value of a single trough level of VPA was questioned by Loiseau et al. (64), based on the observation that the mean fluctuation in concentrations was 113% during a 24-hour period, and consecutive fasting levels were not reproducible.
In conclusion, although VPA levels can be valuable in selected cases, and particularly during combination therapy, a single measurement is often of limited value, and results should be interpreted cautiously (22). Rigid adherence by the physician to the indicated therapeutic range, usually 50 to 100 µg/mL (350 to 700 µmol/L), is not likely to be beneficial to the patient. Saliva VPA levels are of little use because they do not seem to correlate well with plasma levels (41).
Contraindications and Precautions
VPA is contraindicated in patients with hepatic disease, significant hepatic dysfunction, metabolic disorders associated with an increased susceptibility to liver toxicity, or known hypersensitivity to VPA. Box warnings included in the package insert include potentially fatal hepatotoxicity, pancreatitis, and teratogenicity (e.g., spina bifida). The possible adverse effects of VPA are discussed in Chapter 89 and are not reviewed here. However, the risk of some of these
adverse effects can be reduced if certain precautions are taken in patients treated with VPA.
Because infants <3 years old who were taking VPA with other antiepileptic drugs were found to have the highest risk of fatal hepatotoxicity (16), VPA therapy should be preferably avoided in this age group. Benign elevation of liver enzymes can be seen with VPA, and severe hepatotoxicity is usually not preceded by elevation of liver enzymes. Although routine monitoring of liver enzymes during VPA therapy is a common practice, the diagnosis of hepatotoxicity depends mostly on early recognition of the clinical features, which include nausea, vomiting, anorexia, lethargy, and at times loss of seizure control, jaundice, or edema. Similarly, the occurrence of abdominal pain, nausea, vomiting, or anorexia requires evaluation for the possibility of pancreatitis.
Supplementation with L-carnitine may be beneficial in certain patients taking VPA (32,73). Currently, intravenous L-carnitine supplementation is indicated for VPA-induced hepatotoxicity, overdose, or other acute metabolic crisis associated with carnitine deficiency (32). L-Carnitine supplementation is also indicated for the primary plasmalemmal carnitine transporter defect. L-Carnitine supplementation is suggested in the following circumstances: patients with certain secondary carnitine deficiency syndromes, symptomatic VPA-associated hyperammonemia, multiple risk factors for VPA hepatotoxicity or renal-associated syndromes, infants and young children taking VPA, patients with epilepsy using the ketogenic diet who have hypocarnitinemia, patients receiving dialysis, and premature infants who are receiving total parenteral nutrition. The recommended oral L-carnitine dosage is 100 mg/kg/day up to a maximum of 2 g/day. It does not appear that VPA lowers carnitine levels in otherwise healthy and well-nourished children (49). Investigators have suggested that comedication with topiramate represents a risk factor for VPA-induced hypoammonemic encephalopathy (45). Similarly, an association between treatment with VPA and complications of the ketogenic diet has been suggested (4).
One should routinely monitor blood count, including the platelet count, in patients taking VPA. Thrombocytopenia is by far the most frequently diagnosed hematologic adverse effect of VPA, and it is much more common at higher VPA levels (28). Some physicians discontinue VPA therapy before surgical intervention because of VPA-mediated disturbances of hemostasis, although no objective evidence of excessive operative bleeding has been found (2,106,111).
In women, VPA can cause menstrual irregularities (69), hormonal changes, (53, 54, 55), and pubertal arrest (25). One concern has been the increasing recognition of the frequent association between VPA therapy and polycystic ovaries (53, 54, 55,88). Treatment with VPA during the first trimester of pregnancy has been found to be associated with an estimated 1% to 2% risk of neural tube defect in the newborn (9,63,77). Folate supplementation may reduce the risk (107), and a daily dose of at least 1 mg should be recommended to all female patients of childbearing age who are taking VPA. The possible endocrinologic and teratogenic adverse effects of VPA should be discussed with every woman of childbearing age in whom treatment with VPA is being considered.
36a. Ferrie CD, Robinson RO, Knott C, et al. Lamotrigine as an add-on drug in typical absence seizures. Acta Neurol Scand 1995; 91:200-202.
79a. Pisani F, Oteri G, Russo MF, et al. The efficacy of valproate-lamotrigine comedication in refractory complex partial seizures: evidence for a pharmacodynamic interaction.Epilepsia 1999; 40:1141-1146.
81a. Richens A, Davidson DLW, Cartlidge NEF, et al. A multicentre comparative trial of sodium valproate and carbamazepine in adult onset epilepsy. J Neurol Neurosurg Psychiatry 1996;57: 682-687.
85a. Rouan AJ, Meijer JWA, de Beer Pawlikowski N, et al. Valproate-ethosuximide combination therapy for refractory absence seizures. Arch Neurol 1983;40:797-802.