Antiepileptic Drugs, 5th Edition

Phenobarbital and Other Barbiturates


Adverse Effects

Michel Baulac MD*

Joyce A. Cramer BS**

Richard H. Mattson MD***

* Hopital de la Salpetriere, Bat. P. Castaigne, Paris, France

** Associate Research Scientist, Yale University School of Medicine, Veterans Affairs Connecticut Health Care System, West Haven, Connecticut

*** Professor of Neurology and Director of Medical Studies, Department of Neurology, Yale University School of Medicine and Veterans Administration, West Haven, Connecticut

The safety and tolerability profiles of phenobarbital rest on a considerable background of knowledge owing to its worldwide use for almost a century. Phenobarbital has a reputation of safety because serious systemic side effects are very uncommon. Nonetheless, it presents a strong potential for inducing sedation, cognitive impairment, or behavioral disturbances, particularly in children. These types of neurologic adverse effects are not entirely specific to phenobarbital and also can be encountered with most of the other antiepileptic agents. Comparative studies, however, confirming the clinical experience, showed that many of these neurotoxic effects were more frequent and more pronounced with phenobarbital than with the other drugs. Because of this suboptimal tolerability, along with the increasing number of other therapeutic options, use of phenobarbital as long-term antiepileptic treatment continues to decrease in developed countries.


An overview of phenobarbital tolerability is given by the studies in which different antiepileptic drugs have been compared in monotherapy settings. The multicenter Veterans Administration (VA) study (1) has been pivotal from this point of view by showing that the rate of withdrawal for unacceptable adverse effects was 19% with phenobarbital, versus 16% with phenytoin, 12% with carbamazepine, and 33% with primidone. In this double-blind study, therapeutic blood concentrations had to be achieved very rapidly, and, in case of seizure recurrence, daily dosages were increased up to maximum tolerated doses. These aspects of the design may have accounted for a globally high (20%) rate of withdrawal. Two recent, prospective, randomized, but unmasked studies have been conducted in newly diagnosed patients, with designs closer to clinical practice (2,3). These trials allowed a slow titration process and the maintenance of a relatively low dosage if seizures were controlled before therapeutic levels were reached. Globally, in adults (2), 10% of the patients were withdrawn owing to unacceptable adverse effects, with 22% of these subjects randomized to phenobarbital, versus 11%, 5%, and 3% to carbamazepine, sodium valproate, and phenytoin, respectively. A similar study was undertaken in children (3) to compare phenobarbital, carbamazepine, phenytoin, and sodium valproate. Of the 167 children with newly diagnosed epilepsy who were included, 6 of the first 10 children assigned to phenobarbital experienced unacceptable adverse effects, all cognitive or behavioral. No further children were assigned to this drug, leaving the phenobarbital arm of the study uncompleted. These data show that phenobarbital is among the less well tolerated of the conventional or old drugs. There are no comparative data between phenobarbital and any of the recent antiepileptic drugs, some of which are known for a favorable tolerability profile. Nearly all the adverse reactions leading to withdrawal of phenobarbital occurred in the domain of neurotoxicity.


Various neurotoxic adverse effects can be encountered with the long-term use of phenobarbital, even at daily dosages maintaining serum concentrations of the drug in the broad therapeutic range of 15 to 40 µg/mL. Besides changes in affect, behavior, or cognitive function, high serum concentrations may cause neurologic signs of “drunkenness,” including nystagmus, dysarthria, incoordination, and ataxia. Often, the neurotoxic side effects occur together in different degrees. Moreover, the terminologies used for their


description are variable, making it difficult to separate one from the other. The occurrence and the magnitude of these neurotoxic symptoms can be reduced at the initial stages of the treatment by a slow titration process. All these neurologic adverse effects are dose dependent and reversible after dosage reduction, if a reduction is possible in terms of seizure control effectiveness. Treating physicians should be aware that patients or relatives might not report some of these symptoms because they can develop insidiously during long-term use.


The hallmark of barbiturate toxicity in adults is sedation. In the VA study (1), two of three patients complained of sedation at one or more visits in the first year. Interestingly, phenobarbital produced no more acute sedation than the other drugs tested, probably because of cautious dose increases. Complaints of fatigue and tiredness are difficult to quantify and often are variable and subtle. The patient and family may describe listlessness or lack of spontaneity even when excessive sleeping time is not observed. As dosage is increased, overt sleepiness is observable and often manifests as difficulty with arousal in the morning and naps after school or work. An associated loss of interest, particularly in social activity or playing with friends, is common. Butler et al. (4) noted that patients complained of sedation at the onset of treatment when phenobarbital concentrations were only 5 µg/mL. Two weeks later, there were few complaints despite a fivefold increase in the serum levels. Others also have reported that sedation occurred primarily during the first few days of treatment and decreased rapidly as tolerance developed (5,6). Somnolence was even briefer if phenobarbital was restarted after a withdrawal period . It also was found that a dose that caused sedation during initiation of therapy in adults no longer caused sleepiness after 1 or 2 weeks of treatment (5). After tolerance was acquired, major adverse effects of phenobarbital were not observed when the serum concentration was less than 30 µg/mL. A subgroup of 58 patients taking phenobarbital in the VA study was examined at every visit for the first 3 months (1, 2, 4, 8, and 12 weeks) to assess the incidence of acute adverse effects and development of tolerance (7). Of the subgroup studied for tolerance, 33% of patients started on phenobarbital reported initial sedation, declining significantly to 24% by 12 weeks (p < .04). Development of tolerance was evidenced by decreasing symptoms despite increasing phenobarbital concentrations from 18 µg/mL at 2 weeks to 24 µg/mL at 12 weeks.

Mattson et al. (8), however, found many exceptions to the correlation between serum phenobarbital concentrations and complaints of tiredness. The variation among individuals was evident in that some patients were asymptomatic when serum phenobarbital levels were as high as 50 µg/mL, whereas others complained of feeling “drugged” when levels were as low as 15 µg/mL. The usual tolerable range of serum concentrations when phenobarbital is used as the sole drug is 15 to 30 µg/mL.

Neurologic Side Effects

Increasing the dosage of phenobarbital eventually leads to neurologic signs similar to those found with the use of other antiepileptic drugs. Dysarthria, incoordination, ataxia, dizziness, and nystagmus often appear as serum levels exceed 40 µg/mL. The VA study found that at lower levels and at initiation of therapy, these signs and symptoms were significantly less frequent with use of phenobarbital than with carbamazepine, phenytoin, or primidone (p < .03).


Instead of the sedative effect of phenobarbital common in adults, paradoxical effects of the drug in children and, less commonly, in the elderly are insomnia and hyperactivity. Ounsted (9), in reviewing what he called “the hyperkinetic syndrome,” found that many of the children receiving phenobarbital therapy were overactive. The pattern of behavior included signs of distractibility, shortened attention span, fluctuation of mood, and aggressive outbursts. Most of the children were boys. Wolf and Forsythe (10) also found a high incidence of behavioral disturbances. These problems developed in 42% of 109 children receiving daily phenobarbital therapy to prevent recurrence of febrile seizures. Surprisingly, 64% of the children exhibiting hyperactivity had serum phenobarbital concentrations of <15 µg/mL, indicating that such problems can be seen even in what would be considered the low or subtherapeutic range. The authors (10) also suggested that behavioral disturbances associated with phenobarbital use are more likely to become evident in children in the presence of organic brain disease or deficits.

In contrast, Camfield et al. (11) assessed 35 toddlers given phenobarbital and 30 given placebo and found no differences in hyperactivity between the groups after a year. Dose-related irritability and erratic sleep were common in the phenobarbital group, without frank hyperactivity. Reduction from 4 to 5 mg/kg/day to 2 to 3 mg/kg/day resolved these problems in four children. A randomized comparison of phenobarbital and phenytoin intended to detect behavioral side effects showed similar acceptability for the two drugs as monotherapy for childhood epilepsy in rural India (12), but the sensitivity of the procedures used for behavior assessment was limited (13). Another randomized trial comparing phenobarbital with phenytoin and valproate as initial treatment for epilepsy in India did not confirm that hyperactivity problems are more common with phenobarbital, and concluded that this drug should be considered


as a possible first choice only for preschool children (14).

Elderly patients with organic brain disease also may become agitated rather than sedated with use of phenobarbital.

Mood and Affect

Phenobarbital therapy can produce alteration of affect, particularly depression. It is difficult to determine whether such mood changes are a reaction to the often newly diagnosed illness, the addition of another drug to treat severe seizures, or a direct neurotoxic effect of phenobarbital. Clinical observations in children suggest a direct effect of phenobarbital because changes to carbamazepine therapy have been associated with improved mood scores (15). In another study of children (16), phenobarbital (38%) was associated with a higher incidence of depression than was carbamazepine or no treatment (0%). Other studies in adults have not revealed statistically significant changes over time among patients on phenobarbital compared with other drugs (1). Early psychological problems with affect, mood, and cognition were reported by 13% of patients in the VA study at 1 month and by 12% at 3 months (7). This was not significantly different in phenobarbital-treated patients compared with carbamazepine, phenytoin, or primidone treatment. However, 40 of 56 patients treated for at least 1 year reported some psychological effect of phenobarbital during the first year. The risk for depression is increased in patients with a personal or family history of an affective disorder.


A side effect of phenobarbital of considerable potential importance, especially in children, is a possible disturbance in cognitive function. Problems with memory or compromised work and school performance may develop independent of sedation and hyperactivity, although these factors may play a contributory role. Lennox (17) observed a marked impairment in affect and cognitive function in patients whose capacity had already been compromised: “Many physicians in attempting to extinguish seizures only succeed in drowning the finer intellectual processes of their patients.” Such effects often are subtle and difficult to measure despite reports by patients, families, and teachers. Many reports and studies have dealt with this issue of cognitive changes induced by phenobarbital, in adults as in children, but their interpretation sometimes is difficult. Furthermore, some studies have been conducted in children with epilepsy whose persisting seizures or underlying brain damage may interfere with the results, whereas other studies have been done in children who do not have epilepsy who received phenobarbital for the prophylaxis of febrile seizures.

Changes in cognitive function have been measured by various standardized neuropsychological tests. Interestingly, in early reports, institutionalized epileptic patients showed some improvement in intelligence testing after treatment with antiepileptic drugs. Improved test scores could be attributed to decreased seizure frequency or a practice effect from repeated testing. Lennox (17) found 58% of his patients unchanged on subjective evaluation of mentality while using phenobarbital. He separated the improvement of patients because of diminished seizure frequency from the effect of the medication on mentality. A more detailed study by Somerfeld-Ziskind and Ziskind (18) showed no overall change in intelligence quotient (IQ) after phenobarbital therapy for 1 year. Twelve patients actually showed increased IQ scores, whereas 10 patients had lower scores (maximum change, 11 points); 79% had fewer seizures while receiving phenobarbital.

Stores (19) reviewed studies of the effect of phenobarbital on intellectual function in children. He commented that the educational problem for these children appears to be that their attainments fail to match their capacities as measured by standardized tests. Formal studies have not been able to assess this disparity. On careful testing, children treated with phenobarbital can perform at appropriate levels. It is difficult to assess subjective complaints unless the children are treated with a different medication and tested before and after the change from barbiturate therapy.

A double-blind, crossover comparison of psychological and behavioral effects of phenobarbital and valproate was performed in 21 epileptic children (aged 6 to 15 years) by Vining et al. (20). Cognitive function and behavior were significantly diminished during phenobarbital therapy (p < .01), although differences were subtle. Overall intelligence assessed by the Wechsler Intelligence Scale for Children (Revised) (WISC-R) showed significantly lower performance and full-scale IQ scores for phenobarbital than for valproate treatment periods. Differences were seen both in verbal and nonverbal tasks, particularly for complex tests. The extensive neuropsychological testing showed important problems with epilepsy and developmental problems in children (20). The comparison with valproate, a drug not considered likely to cause cognitive impairments, provided information suggesting that phenobarbital can affect childhood learning and behavior.

In another study, children with epilepsy treated with phenobarbital (n = 32) or valproate (n = 32) were compared with healthy children (n = 66) for WISC scores. Total, verbal, and performance scores were lower for children receiving phenobarbital than for control subjects (21). Although valproate-treated and healthy children demonstrated a learning effect with improved scores on retest, the phenobarbital group did not improve, suggesting impaired ability to learn.

Camfield et al. (22) found a trend toward decreased memory and concentration scores in epileptic children taking


phenobarbital compared with a placebo group (p < .07) that correlated well with serum levels (p < .05). They suggested caution in long-term exposure of children to phenobarbital, particularly at high dosage. In a careful study of children receiving an average daily dosage of 1.8 mg/kg of phenobarbital, however, Wapner et al. (23) compared learning behavior and intelligence before therapy and 6 weeks later and found that phenobarbital did not affect the function of the children in the classroom situation. Although seizure control was incomplete, there was no significant change in learning or intellect compared with a control group. It is possible that there is considerable variability in susceptibility to the adverse effects of barbiturates on learning and cognition. Recent case reports suggest that the manifestations of phenobarbital toxicity in children may even include regression of developmental milestones, mimicking a neurodegenerative disorder (24).

Schain et al. (15) found that when carbamazepine was substituted for phenobarbital in children, several mental functions were improved. In particular, they found a statistically significant difference in intelligence (as measured by WISC) and results of three problem-solving tests of attentiveness and impulse control. Parents and teachers also reported a significant improvement in alertness and attentiveness. These drug changes improved seizure control, but the psychological improvements were considered to be a function of removal of the sedating drug.

Other investigators (6) studied adults who had received phenobarbital for 2 weeks, allowing for the partial development of tolerance before performance testing. They found that phenobarbital did not diminish performance on simple tasks requiring attention but did affect tasks requiring sustained effort. Even the tasks requiring sustained effort showed improvement when patients were stimulated during the testing. Others also showed the difference between self-paced tests and tests in which sustained attention was necessary (25). Impaired vigilance and sensory perception during phenobarbital use have been noted in patients of average intelligence.

Hutt et al. (6) tested the effects of phenobarbital after tolerance had developed. Although sedation was less at the time of testing compared with the early acute effect, performance on perceptual-motor tests was significantly impaired in proportion to serum phenobarbital concentration. Tests requiring sustained vigilance were affected negatively. The author defined several factors that were significantly correlated with test performance: (a) serum phenobarbital concentration; (b) difficulty and duration of the task; and (c) tester interaction with the subject (i.e., external stimulation).

A closely related but separate issue is the question of memory impairment, which is a common complaint from epileptic patients and which unquestionably is related, in part, to the brain lesion (26). In detailed studies of patients tested when phenobarbital concentrations were at moderate and then high therapeutic levels, MacLeod et al. (27) compared short-term versus long-term memory storage. They found short-term memory scanning significantly impaired when phenobarbital levels were high, but retrieval of information stored in long-term memory was undiminished. Although this study was unfortunately brief in its 1-week trial at each dose, the data suggest that phenobarbital impairs access to information in short-term but not longterm memory. The authors suggested that impaired short-term memory might be an important influence in acquisition of new information because of impaired attention span. Oxley (28) indicated that a significant improvement in memory function could be achieved after a reduction in phenobarbital dose. This report was of interest because the patients experienced increased seizure frequency when the barbiturate level dropped, indicating that it is not seizure activity that impairs memory. In summary, the assessment indicated that phenobarbital has a deleterious effect on short-term memory, with test performance related to dose. Even when the serum concentration is within the therapeutic range, ability to concentrate and perform simple tasks is reduced.

Several studies have been done in children who received phenobarbital for the prevention of febrile seizures. Hirtz et al. (29) compared IQs of children receiving phenobarbital after a febrile seizure with an untreated group. IQs were significantly lower (7 points) in the phenobarbital-treated group, remaining 5 points lower even after medication was discontinued. The authors concluded that treatment with phenobarbital depressed cognitive performance in children. Farwell et al. (30) reported on the effect of phenobarbital on intelligence of children receiving the medication as prophylaxis for febrile seizures. Two years after randomization to phenobarbital therapy, children assigned to that group had Stanford-Binet IQ scores 8.4 points lower than children in the group assigned to placebo (p < .006). IQ scores remained 5.2 points lower for the group assigned to phenobarbital when tested 6 months after that group had treatment discontinued (p < .05). Only 64% of children continued to receive phenobarbital throughout the study, and only two-thirds of the phenobarbital blood levels were above 15 µg/mL during follow-up. These children were retested 3 to 5 years later, after they had entered school, to determine whether those effects persisted over the longer term and whether later school performance might be affected (31). On follow-up testing of 139 (of the original 217) children who had experienced febrile seizures, the phenobarbital group scored significantly lower than the placebo group on the Wide Range Achievement Test, and a nonsignificant mean difference of 3.71 IQ points was observed on the Stanford-Binet, with the phenobarbital-treated group scoring lower. The authors concluded that there might be a long-term adverse cognitive effect of phenobarbital on the developmental skills (language/verbal) being acquired during the period of treatment.



Another prospective study was performed to ascertain whether febrile convulsions in early childhood are associated with neurocognitive attention deficits in school-age children (32). A total of 103 children, confirmed to have febrile seizures by the age of 3 years, were followed up until at least the age of 6 years. An achievement test, behavioral ratings, and computerized neurocognitive battery assessing various subcomponents of attention were given to 87 children with febrile seizures and 87 randomly selected population-matched control subjects. Attention performance of the two groups was comparable, including the subgroup of children treated with phenobarbital.

Libido and Potency

A side effect of phenobarbital not documented in the literature and too little appreciated by treating physicians is decreased libido and impotence (1). In clinical practice, we have found that responses to specific questions reveal numerous complaints from men receiving phenobarbital. It is difficult to assess whether the problem is organic or related to psychological depression. Occasionally, dosage reduction improves the problem. Fifteen percent of the patients included in the VA study complained of decreased libido or potency, and this problem was found to be more common in patients treated with phenobarbital or primidone than in those receiving carbamazepine or phenytoin (p < .06) (1). Fourteen percent of 56 patients treated for 1 year with phenobarbital reported a transient or continuous decrease in sexual function. The reports increased over time, indicating that this is neither an acute problem nor one for which tolerance develops. Lowering the dosage allowed improvement in some instances, but the drug was discontinued for patients who did not improve.

Multiple mechanisms, including altered metabolism of testosterone (33,34) and psychosocial factors (35), may be responsible for these complaints. However, the problem usually disappeared when carbamazepine or phenytoin was substituted for phenobarbital, but not when phenobarbital was changed to another barbiturate. Psychosocial factors were comparable, testosterone levels (both total and free) were equal, and enzyme-inducing properties are similar for all drugs in the four treatment groups. Consequently, we concluded that the changes in sex behavior might be a direct neurogenic effect.

Precaution and Management

Many neurotoxic side effects improve with a simple reduction in dosage. Of course, improvement is gradual because of the slow elimination of phenobarbital. Such lowering of serum concentration provides less protection against seizures (36). In the past, when seizure control could be achieved only at the cost of neurotoxic side effects, sedation or hyperactivity sometimes were ameliorated with concomitant administration of amphetamines. Today, it is less necessary to subject patients to the additional complications attendant on the use of these stimulants. A change to treatment with an alternative antiepileptic drug may be equally effective and spare some side effects.

Dependence, Habituation, and Withdrawal

Phenobarbital shares the properties of other barbiturates in that prolonged use produces dependence, and abrupt discontinuation after high dosage produces abstinence symptoms. Such symptoms include anxiety, emotional lability, insomnia, tremors, diaphoresis, confusion, seizures, and possible status epilepticus (37,38). Reinstituting the drug can reverse these symptoms. It has been suggested that discontinuation of phenobarbital in epileptic patients may lead to exacerbation of seizures not only because of the underlying epilepsy but because of an additional barbiturate withdrawal mechanism (5). Even with abrupt discontinuation, the slow elimination of phenobarbital results in slowly decreasing plasma drug levels. Even so, gradual tapering may be advisable. In a study of the impact of withdrawal of different antiepileptic drugs on seizure recurrence (39), 1,013 patients, in remission of epilepsy for at least 2 years, were randomized to continued therapy or slow withdrawal over 6 months and were followed up for a median period of 5 years. No evidence was found that withdrawal of phenobarbital was associated with withdrawal seizures.

Because phenobarbital can cross the placenta and enter the fetal system, special care must be taken during the neonatal period of children born to mothers who received phenobarbital. The neonatal withdrawal syndrome was described by Desmond et al. (40) for infants born to epileptic mothers. The infants were allowed to withdraw from phenobarbital postpartum. Hyperexcitability, tremor, irritability, and gastrointestinal upset continued for several days to several months. Although the withdrawal syndrome is similar among infants born to heroin addicts, barbiturate addicts, and epileptic mothers, the infants of women on antiepileptic doses of phenobarbital have a milder and briefer withdrawal experience, with good results for all infants (41,42). There is no apparent residual damage after withdrawal. To calculate the probable length of withdrawal in neonates, it should be noted that the phenobarbital concentration in umbilical cord serum is approximately equal to the maternal serum concentration (43). The rate of elimination of phenobarbital in neonates often is slower than that in adults.

Systemic Toxicity

Megaloblastic Anemia

Megaloblastic anemia has been described during treatment with phenobarbital alone or, more commonly, when it is used with other antiepileptic drugs, particularly phenytoin.


Anticonvulsant megaloblastic anemia probably occurs in less than 1% of patients; the incidence was 0.15% to 0.75% in one report (44). The etiology and pathogenesis of macrocytosis and megaloblastic anemia during antiepileptic drug therapy are unknown, but these conditions usually respond to folate therapy.

Folate Deficiency

Frank serum and red blood cell folate deficiency is relatively common. Reynolds (45) surveyed 16 reports in which from 27% to as high as 91% of patients had subnormal serum folate levels, averaging 52%, in patients receiving long-term therapy with phenytoin, phenobarbital, or primidone. The significance of low folate levels is controversial.

Reynolds and Travers (46) reported improvement in psychiatric abnormalities in patients whose low serum folate concentrations were treated with folate therapy. However, such subjective observations are difficult to assess. Although controlled trials have not confirmed that replacement folate therapy in patients receiving phenobarbital or phenytoin either aggravates seizure susceptibility or improves patients' psychological status (47), it is possible that in some patients administration of folate therapy exacerbates seizures (45). Mattson et al. (48) found that serum phenobarbital and phenytoin concentrations decreased when folic acid was given in very high doses. It is possible that reports of seizure exacerbation resulted in part from the decrease in drug concentration rather than from an epileptogenic activity of folate, although the mechanism of this interaction is unknown. Although an inverse correlation exists between folate and phenobarbital levels in both serum and cerebrospinal fluid (CSF) (49), Mattson et al. (48) found no change in CSF folate concentration during folic acid therapy. In fact, animal studies (50) show clearly that even in severe folate deprivation, the brain maintains sufficient folate.

The significance of folate deficiency remains speculative. Other than in cases of obvious megaloblastic anemia, subnormal serum folate probably requires no therapeutic intervention. Except perhaps during pregnancy, proper nutritional balance is sufficient to maintain adequate folate levels during antiepileptic drug therapy. A clear exception exists for women of childbearing age. Supplementary folate reduces the risk of neural tube defects in the developing fetus and should be prescribed before conception (51).

Vitamin K

Another hematologic abnormality caused by antiepileptic drug therapy affects vitamin K. Phenobarbital and phenytoin, which enter the liver of the fetus, can interfere with vitamin K and decrease production of vitamin K-dependent clotting factors. This can occur even in the presence of normal clotting factors in mothers receiving drug therapy. Mountain et al. (52) reported 7 neonates with a severe coagulation defect in a series of 16 neonates whose mothers received various antiepileptic drugs (including 13 receiving barbiturates). The neonate can sustain intraperitoneal, intrathoracic, or intracranial bleeding if vitamin K-dependent coagulation factors are deficient. These signs occur within the first day or two postpartum. Vitamin K administered to mothers prepartum or to neonates at the time of birth prevents this coagulation deficiency (52,53).

Bone Disorders

Antiepileptic drug therapy may affect calcium and vitamin D metabolism, leading to hypocalcemia or, rarely, osteomalacia (54). Despite a high incidence of subnormal calcium levels, no more than 10% of epileptic patients were found to have osteomalacia, with the disorder developing only after many years of drug therapy (55). The incidence of this disorder may relate to climate and life-style (i.e., lack of exposure to sunlight) and diet. Bone mineral density measurements may help to detect subtle bone loss in children receiving phenobarbital over 24 months (56). Induction of liver enzymes leading to increased hydroxylation of vitamin D is a probable mechanism for altered calcium metabolism (57). Reversing signs of deficiency can be accomplished with less than 125 µg of vitamin D3 per week (58).

Cardiovascular Risks

The potential effects of long-term phenobarbital therapy on the risk of atherosclerosis-related diseases have been differently addressed in the literature. Epidemiologic surveys suggest that antiepileptic drugs may have a protective effect against cardiovascular disease, with 29% less mortality due to ischemic heart disease than in respective control subjects (59). This also was supported by certain studies of serum lipoprotein patterns: through the activation of liver microsomal function, phenobarbital induces apolipoprotein A-1 and high-density lipoprotein synthesis and increases their serum levels, which would be expected to protect against atherosclerosis (60,61)

Other results suggest that, in contrast to the aforementioned finds, the effects on the serum lipid profile of longterm treatment with hepatic enzyme-inducing antiepileptic drugs (e.g., carbamazepine and phenobarbital) perhaps are not beneficial with regard to the risk of atherosclerosis-related disease: elevations of lipoprotein (1), an independent risk factor for atherosclerosis, have been observed in adults (62) and in children (63), and decreased levels of apolipoprotein A have been found in children (64). Finally, an elevated plasma concentration of homocysteine, another established risk factor for atherosclerosis, seems to be associated with antiepileptic drug treatments, including phenobarbital (65). Monitoring of serum cholesterol may be recommended in patients receiving long-term phenobarbital treatment, whatever their age.



Connective Tissue Disorders

A higher incidence of Dupuytren's contractures, palmar nodules, frozen shoulder, Ledderhose's syndrome (plantar fibromatosis), Peyronie's disease, heel and knuckle pads, and general joint pain has been noted in patients taking antiepileptic drugs than in the general population (66,67). These connective tissue disorders were first linked to patients with epilepsy in 1925, when Maillard and Renard (68) called attention to joint pain associated with the use of the newly introduced barbiturates, especially phenobarbital. Soon after, Beriel and Barbier (69) termed this disorder rheumatism gardenalique. Until recently, conflicting evidence was available to differentiate among probable etiologic agents. Data from another VA study (70) provided evidence of a statistically significant association between use of phenobarbital and primidone for at least 6 months and onset of all 10 cases of connective tissue disorders (71). None of the 107 patients receiving carbamazepine alone or the 121 on phenytoin experienced a problem (p < .001). Critchley et al. (71) were able to define phenobarbital as a common cause in contractures, and Janz and Piltz (72) associated primidone with frozen shoulder.

The incidence of barbiturate-related disorders ranges from 5% to 38%, depending on the population studied. Froscher and Hoffman (73) noted that their general outpatients with epilepsy had a 5% incidence of contractures, similar to that found in the VA study outpatients, but they noted a higher incidence (20%) in patients with severe epilepsy. Noble (74) reported a surgeon's point of view, having seen Dupuytren's contractures in 10% to 38% of institutionalized patients with epilepsy. Duration of treatment probably increases the incidence of disorders.

Reversibility has been seen during continued drug use (72), but improvement is most likely when the barbiturate is stopped early, particularly in patients with barbiturateassociated frozen shoulder and joint pain (67,75). Mattson et al. (70) also described clearing of signs and symptoms when carbamazepine, phenytoin, or valproate was substituted for the barbiturate.

Hepatic and Metabolic Disorders

Phenobarbital is a hepatotoxin only in unusually susceptible individuals. Liver disease induced by antiepileptic drugs, particularly phenobarbital or phenytoin, appears not to be dose dependent and has a low incidence. Most drugs are indirect hepatotoxins, selectively blocking metabolic pathways and producing structural changes by precise biochemical lesions (76). Idiosyncratic acute hepatic injury may be cytotoxic, cholestatic, or mixed. Cytotoxicity can lead to liver necrosis or cholestasis. These drugs probably produce hepatocellular injury—for example, liver necrosis or cholestasis (76) (see section on Hypersensitivity Reactions, later).

Enzymatic Effects

Antiepileptic drugs, particularly phenobarbital, are potent inducers of hepatic microsomal enzymes, which can lead to enhanced metabolism of other drugs or endogenous substances (77). Although some of these effects are considered drug interactions, the basis of these interactions must be considered a hepatic or metabolic side effect of phenobarbital therapy. When metabolism of other compounds is accelerated, the end effect of drugs or substances can be diminished or negated, or pathways can be modified to produce different, potentially effective or toxic metabolites. For example, the induction of microsomal valproate metabolism increases the concentration of the 4-en VPA hepatotoxic metabolite (78). Phenobarbital stimulates hydroxylation pathways related to numerous endogenous and exogenous substances, including thyroid hormones, sex hormones, and other steroids (79). Phenobarbital increases the excretion of 6β-hydroxycortisol, leading to decreased plasma cortisol half-life (80). It also has been shown to increase the rate of metabolism of dexamethasone and prednisone. The resulting lower serum concentrations of these drugs used by patients with bronchial asthma or rheumatoid arthritis disturbed the treatment of their underlying disorder (79). Withdrawal of phenobarbital allowed these changes to reverse (81). Enhanced hormone metabolism can cause failure of oral contraceptives, particularly with low-dose pills (33,82). Interference with the anticoagulation activity of coumarin drugs has been related to phenobarbital therapy (83). Erratic control of anticoagulation with decreased prothrombin time was noted during phenobarbital therapy, and increased dosage of the anticoagulant drug was necessary. However, if phenobarbital is withdrawn, allowing decreased enzyme stimulation, the other drug also requires dosage reduction, or bleeding may occur.

Patients treated with phenobarbital often show increased levels of serum alkaline phophatase and γ-glutamyl transferase. These abnormalities represent an epiphenomenon of enzyme induction and have no clinical significance (79).


Because phenobarbital can induce synthesis of liver enzymes, it has been shown to enhance the synthesis of γ-aminolevulinic acid (ALA) synthetase, which can aggravate porphyria. Granick (84) hypothesized that drugs such as barbiturates may interact with heme, thereby diminishing inhibition of enzymes controlling ALA synthase production. Hereditary acute porphyria can be exacerbated when barbiturates are used.


Bilirubin glucuronidation is induced by phenobarbital, and this has been used to treat neonatal hyperbilirubinemia (85).

Ethanol and Other Drugs

Cramer and Scheyer (86) reviewed the interactions between ethanol and antiepileptic drugs, noting several characteristics


shared by phenobarbital and ethanol. Both compounds lead to hypertrophy of hepatic smooth endoplasmic reticulum, inducing nonspecific increases in numerous hepatic drug-metabolizing enzymes. Both compounds are oxidized by nicotinamide-adenine dinucleotide hydrogenase (NADH) microsomal systems (87,88). Barbiturate-hydroxylating enzymes are increased in men given alcohol (89). Enzyme induction allows for increased clearance of drugs in alcoholic patients as well as in those receiving barbiturates. Phenobarbital used with other drugs of abuse significantly affects their metabolism. In addition to altering ethanol metabolism, phenobarbital increases the rate of heroin deacetylation. This increase in detoxification is dose related, parallel to increased enzyme induction (90). Conversely, barbiturates can reduce alcohol dehydrogenase activity, allowing high levels of alcohol to occur when both compounds are used concurrently. The synergy of barbiturate and ethanol toxicity can cause respiratory depression, leading to unexpected death (91).

Gastrointestinal Disturbances

In the VA study (7), early gastrointestinal complaints occurred in 2% to 3% of patients, compared with 7% to 18% of patients on carbamazepine, phenytoin, or primidone (7). Although probably of central origin when nausea and vomiting occur acutely, as with primidone, some local gastrointestinal irritation is possible.



Animal studies have shown liver tumors appearing with the use of phenobarbital and other drugs that activate liver enzymes. Although enzyme induction causes an increase in liver size, it also may protect against the carcinogenicity of other compounds (i.e., known chemical carcinogens) by enhancing their metabolism.

There is no evidence of an increased frequency of liver tumors in patients taking phenobarbital. In fact, Clemmensen et al. (92) found a decrease in tumor incidence in patients receiving anticonvulsants. White et al. (93) found an increase in cancer deaths for epileptic patients, but this was not statistically significant. A Danish study reviewed the incidence of cancer among 8,004 hospitalized patients with epilepsy. Their data suggest that exposure to phenobarbital did not correlate with brain tumor incidence, and that phenobarbital is not a carcinogen for humans (94).

Hypersensitivity Reactions

Phenobarbital causes various types of skin reactions. These usually are mild maculopapular, morbilliform, or scarlatiniform rashes that fade rapidly when drug administration is stopped. The incidence has been reported to be as low as 1% to 3% of all patients receiving barbiturates (95). Overall, a hypersensitivity reaction to phenobarbital occurred in 9% of patients in the VA study (7), not significantly different from rate of the reaction to carbamazepine, phenytoin, or primidone. The rash was transient and did not require a change in treatment in 5 of 13 patients. None required hospitalization. Considering the universal usage of this drug, reports of exfoliative dermatitis, erythema multiforme, Stevens-Johnson syndrome, or toxic epidermal necrolysis are impressively rare. Welton (96) reported a case of exfoliative dermatitis with hepatitis caused by phenobarbital.

Hypersensitivity reactions are characterized by rash, eosinophilia, liver toxicity, lymphadenopathy, and fever, and may occur with phenobarbital as with other antiepileptic drugs (97). Histologic changes in the liver show eosinophilic or granulomatous inflammation (97). McGeachy and Bleemer (98) reviewed 17 instances of fatal sensitivity to phenobarbital. Another half-dozen cases of acute reaction to barbiturates and details of treatment were reported by Yatzidis (99). Corticosteroids may be of value in treatment (98,100), although their use in severe reactions affecting the skin has been questioned (101). Once sensitivity has been documented, only rarely should the patient be reexposed to the barbiturate (100).

Systemic lupus erythematosus (SLE) can develop with use of antiepileptic drugs. Alarcon-Segovia (102) suggests that the drugs elicit production of antinuclear antibodies by altering nuclear components. This may unmask SLE in predisposed individuals and can be reversed by prompt discontinuation of the drug.


Long-Term Developmental Effects

Schain and Watanabe (103) reported that young rats showed retardation of brain growth and changes in behavior after long-term administration of phenobarbital. Hiilesmaa et al. (104) subsequently reported decreased fetal head growth associated with maternal use of antiepileptic drugs and suggested a phenobarbital effect. Several groups have found disturbed neuronal development in cultures containing phenobarbital comparable with or greater than what is found with other antiepileptic drugs (105, 106, 107). These findings may not be directly applicable to human use, but they raise special concerns because phenobarbital sometimes is used in pregnancy and in the treatment of neonates.

The neuropsychological consequences of antiepileptic treatment during pregnancy for school-age children and adolescents have been evaluated in 67 subjects (108). Maternal epilepsy and phenobarbital therapy during pregnancy appeared to have long-term effects on the offspring well into adolescence, as evidenced in electroencephalographic


patterns, minor neurologic dysfunction, and intellectual performance. Severity of effects was most marked in the polytherapy group. Similar consequences were suggested in adults. A Danish study included 114 adult men born between 1959 and 1961 who were exposed to phenobarbital during gestation through maternal medical treatment and whose mothers had no history of a central nervous system disorder and no treatment during pregnancy with any other psychopharmacologic drug. Subjects exposed prenatally to phenobarbital had significantly lower verbal intelligence scores (approximately 0.5 standard deviation) than predicted (109). All these findings should be interpreted cautiously because of possible selection bias and influence of confounders. Moreover, detrimental environmental conditions may interact with prenatal biologic insults to magnify neuropsychological dysfunctions.


With the current awareness that most antiepileptic drugs have some teratogenic effects, counseling for the patient who wishes to become pregnant is indicated. It is difficult to correlate specific teratogenic effects with individual drug use in clinical studies of teratogenicity because of the frequent use of multiple drugs as well as other independent environmental and genetic risk factors. Animal studies suggest that most antiepileptic drugs have some potential for teratogenicity. However, when Chatot et al. (110) grew rat embryos in a medium containing human serum from patients taking antiepileptic drugs, malformed embryos were less common for phenobarbital than for carbamazepine, phenytoin, or valproate.

Speidel and Meadow (111) reported that a variety of malformations occurred in the offspring of six women who used only phenobarbital during their pregnancy: tracheoesophageal fistula, ileal atresia, diaphragmatic hernia with pulmonary hypoplasia, thumb and radius aplasia, congenital heart lesion with microcephaly, mental retardation, hypospadias, and meningomyelocele. In this report and in a similar retrospective survey by Nelson and Forfar (112), the lack of characteristic malformation associated with sole phenobarbital use suggests coincidence rather than causal relationship.

Fedrick (113) found phenytoin far more teratogenic than phenobarbital, but when the two drugs were used in combination, teratogenicity was even more pronounced. Only 4.9% of infants born to mothers who received only phenobarbital during the first trimester were known to have birth defects. This was much lower than incidence rates of 15.2% for sole phenytoin therapy and 22% for combined. Surprisingly, in the same study there was a 10.5% incidence of malformation when mothers with epilepsy took no drugs during pregnancy. A large cooperative study in the United States and Finland (114) implicated phenytoin as a possible teratogen, but the question also was raised as to whether fetal damage was attributable to antiepileptic drugs or to epilepsy itself.

The Italian Multicenter Cohort Study (115) found reduced birth weight and head circumference in offspring of women treated with phenobarbital compared with other drugs. However, children of untreated epileptic women had the same outcome. The major drawback in all of these studies is the difficulty in obtaining precise information about maternal drug use. There is some evidence that higher drug dose and serum concentration correlate with increasing risk of malformation (116).

A prospective French study followed 156 pregnancies in women with epilepsy, 72 of whom were receiving phenobarbital alone (117). Valproate plus phenobarbital was the most common combination therapy. Microcephaly occurred in 18% of births in the valproate-only group and in 27% of the valproate plus phenobarbital group. The investigators also found a statistical association between cardiac defects and combined use of phenobarbital plus phenytoin, although phenytoin was the only significant variable in multivariate analyses.

The risk of major congenital abnormalities associated with specific antiepileptic drug regimens was assessed in a large retrospective cohort (118). The study included 1,411 children born between 1972 and 1992 in four provinces in The Netherlands who were born to mothers with epilepsy and using antiepileptic drugs during the first trimester of pregnancy, and 2,000 nonepileptic matched control subjects. Significantly increased risks of major congenital abnormalities for carbamazepine and valproate monotherapy were found, with evidence for a significant dose-response relationship for valproate. The risk of major congenital abnormalities was nonsignificantly increased for phenobarbital monotherapy when caffeine comedication was excluded, but a significant increase in risk was found when caffeine was included.

In a prospective, case-control cohort study of 174 pregnant women with epilepsy and their offspring (119), outcomes were compared with those of a control group of 355 healthy women and their offspring. Anomalies and fetal death were the primary outcome measures. Abnormal outcomes were associated with three major antiepileptic drugs (carbamazepine, phenytoin, and phenobarbital). In terms of abnormal outcome (death and anomalies), phenobarbital was associated with the highest relative risk, phenytoin with an intermediate relative risk, and carbamazepine with the lowest relative risk.

An analysis of pooled data from five prospective European studies found that prenatal exposure to either carbamazepine or valproate monotherapy was associated with a fivefold increase in risk of congenital malformations compared with the offspring of untreated nonepileptic mothers. Among 1,221 pregnancies exposed to antiepileptic drugs, the risk of malformations (relative to offspring exposed to


phenytoin monotherapy) was increased for the combination of phenobarbital and ethosuximide (relative risk, 9.8), and the combination of phenytoin, phenobarbital, carbamazepine, and valproate (relative risk, 11.0).

A large, prospective, single-center study found that the rate of malformations in the offspring of 517 mothers exposed to antiepileptic drugs during pregnancy was 9.7%, without a significant difference in rates between exposure to polytherapy and exposure to monotherapy (9.6% versus 10.%, respectively). Among monotherapy exposures, structural abnormalities occurred in 15.9% of exposures for valproate (n = 44), compared with 8.6% for primidone (n = 35), 7.1% for carbamazepine (n = 113), 4.8% for phenobarbital (n = 83), and only 3.2% on phenytoin (n = 31). There were no malformations in the offspring of 25 untreated patients.

In another recent prospective analysis of 983 offspring of mothers with epilepsy (120), the incidence of congenital malformations in offspring without drug exposure was 3.1%, versus an incidence with drug exposure of 9.0%. The highest incidence in offspring exposed to a single antiepileptic drug occurred with primidone (14.3%), which was followed by valproate (11.1%), phenytoin (9.1%), carbamazepine (5.7%), and phenobarbital (5.1%).

In view of these data, which sometimes are contradictory, phenobarbital cannot be presented as a safe option for antiepileptic treatment during pregnancy. Drug selection may be influenced by the comparative risks carried by the other options on the one hand, and by other factors, including the need for using the agent that provides the best seizure control, on the other hand.


Frank overdose of phenobarbital causing serum levels in excess of 50 to 60 µg/mL leads to progressive neurologic dysfunction and depression in levels of consciousness, sometimes even in patients on long-term therapy. Excessively high doses first produce ataxia, dysarthria, nystagmus, incoordination, and uncontrollable sleepiness. As the serum levels rise, these effects progress to stupor and coma. Ultimately, depression of cardiorespiratory function may lead to death. The severity of central nervous system depression is much greater in the drug-naive patient, in whom a level of 80 µg/mL is considered potentially lethal (121). Because of tolerance, the occasional individual on prolonged therapy may remain almost unaffected by serum levels that cause unconsciousness in the naive individual. Nonetheless, concentrations above 70 µg/mL can be expected to compromise levels of consciousness in almost all individuals.

In cases where reversing side effects and lowering serum concentration must be rapid, elimination can be accelerated by alkalinization and induction of forced diuresis (4). Approximately one-third of an oral dose of phenobarbital is found in the urine unchanged. When necessary, administration of parenteral fluids up to 5 mg/kg/hr can increase excretion severalfold (122). If some of the fluid given is 1.25% sodium bicarbonate, the alkalinization of blood and urine further enhances elimination after overdose. At pH 7.4, 60% of phenobarbital is ionized and crosses cellular membranes poorly. Penetration into and out of tissue is possible for the 40% of the drug that is nonionized. When acidosis occurs, a higher percentage of phenobarbital is nonionized, allowing passage into the intercellular space. This effectively increases tissue concentrations without any change in total-body phenobarbital. Alkalosis has an opposite effect and leads to movement of phenobarbital out of brain and other tissues. Similarly, alkalinization of urine can appreciably increase elimination of phenobarbital from the body. Activated charcoal, by binding enterally secreted phenobarbital, also is effective in enhancing the elimination of the drug from the body, even when it is administered in the postabsorptive phase (123).


  1. Mattson RH, Cramer JA, Collins JF, et al. Comparison of carbamazepine, phenobarbital, phenytoin, and primidone in partial and secondary generalized tonic-clonic seizures. N Engl J Med1985;313:145-151.
  2. Heller AJ, Chesterman P, Elwes RD, et al. Phenobarbitone, phenytoin, carbamazepine, or sodium valproate for newly diagnosed adult epilepsy: a randomised comparative monotherapy trial. J Neurol Neurosurg Psychiatry1995;58:44-50.
  3. de Silva M, MacArdle B, McGowan M, et al. Randomised comparative monotherapy trial of phenobarbitone, phenytoin, carbamazepine, or sodium valproate for newly diagnosed childhood epilepsy. Lancet1996;347:709-713.
  4. Butler TC, Mahafee C, Waddell WJ. Studies of elimination accumulation, tolerance and dosage schedules. J Pharmacol Exp Ther1954;111:425-435.
  5. Buchthal F, Svensmark O, Simonsen H. Relation of EEG and seizures to phenobarbital in serum. Arch Neurol1968;19: 567-572.
  6. Hutt SJ, Jackson PM, Belsham A, et al. Perceptual motor behaviour in relation to blood phenobarbitone level: a preliminary report. Dev Med Child Neurol1968;10:626-632.
  7. Mattson RH, Cramer JA, Collins JF, and the VA Epilepsy Cooperative Study Group. Early tolerance to antiepileptic drug side effects: a controlled trial of 247 patients. Recent Adv Epilepsy1986;149-156.
  8. Mattson RH, Williamson PD, Hanahan E. Eterobarb therapy in epilepsy. Neurology1976;26:1014-1017.
  9. Ounsted C. The hyperkinetic syndrome in epileptic children. Lancet1955;1:303-311.
  10. Wolf SM, Forsythe A. Psychology, pharmacotherapy and new diagnostic approaches. Adv Epileptol1977; 124-127.
  11. Camfield CS, Chaplin S, Doyle AB, et al. Side effects of phenobarbital in toddlers: behavioral and cognitive aspects. J Pediatr1979;95:361-365.
  12. Pal DK, Das T, Chaudhury G, et al. Randomised controlled trial to assess acceptability of phenobarbital for childhood epilepsy in rural India. Lancet1998;351:19-23.



  1. Trevathan E, Medina MT, Madrid A. Antiepileptic drugs in developing countries. Lancet1998;351:1210-1211.
  2. Thilothammal N, Banu K, Ratnam RS. Comparison of phenobarbitone, phenytoin with sodium valproate: randomized, double-blind study. Indian Pediatr1996;33:549-555.
  3. Schain RJ, Ward JW, Guthrie D. Carbamazepine as an anticonvulsant in children. Neurology1977;27:476-480.
  4. Brent DA, Crumrine PK, Varma R, et al. Phenobarbital treatment and major depressive disorder in children with epilepsy: a naturalistic follow-up. Pediatrics1990;85:1086-1091.
  5. Lennox WG. Brain injury, drugs and environment as causes of mental decay in epilepsy. Am J Psychiatry1942;99:174-180.
  6. Somerfeld-Ziskind E, Ziskind E. Effect of phenobarbital on the mentality of epileptic patients. Arch Neurol Psychol1940;43: 70-79.
  7. Stores G. Behavioral effects of antiepileptic drugs. Dev Med Child Neurol1975;17:647-658.
  8. Vining EPG, Mellits ED, Dorsen MM, et al. Psychologic and behavioral effects of antiepileptic drugs in children: a doubleblind comparison between phenobarbital and valproic acid. Pediatrics1987;80:165-174.
  9. Calandre EP, Dominguez-Granados R, Gomez-Rubio M, et al. Cognitive effects of long-term treatment with phenobarbital and valproic acid in school children. Acta Neurol Scand1990; 81:504-506.
  10. Camfield CS, Chaplin S, Doyle AB, et al. Side effects of phenobarbital in toddlers: behavioral and cognitive aspects. J Pediatr1979;95:361-365.
  11. Wapner I, Thurston DL, Holowach J. Phenobarbital: its effect on learning in epileptic children. JAMA1962;182:937.
  12. Shinnar S, Kang H. Idiosyncratic phenobarbital toxicity mimicking a neurodegenerative disorder. J Epilepsy1996;7:36-37.
  13. Kornetsky C, Orzack MH. A research note on some of the critical factors on the dissimilar effects of chlorpromazine and secobarbital on the digit symbol substitution and continuous performance tests. Psychopharmacology1964;6:79-86.
  14. Delaney RC, Rosen AJ, Mattson RH, et al. Memory function in focal epilepsy: a comparison of nonsurgical, unilateral temporal lobe and frontal lobe samples. Cortex1980;16: 103-117.
  15. MacLeod CM, Dekaban AS, Hunt E. Memory impairment in epileptic patients: selective effects of phenobarbital concentration. Science1978;202:1102-1104.
  16. Oxley J. The effect of antiepileptic drugs on psychological performance. Epilepsy Int1979(abstr).
  17. Hirtz DG, Lee YJ, Ellenberg JH, et al. Survey on the management of febrile seizures. Am J Dis Child1986;140:909-914.
  18. Farwell JR, Lee YJ, Hirtz DG, et al. Phenobarbital for febrile seizures: effects on intelligence and on seizure recurrence. N Engl J Med1990;322:364-369.
  19. Sulzbacher S, Farwell JR, Temkin N, et al. Late cognitive effects of early treatment with phenobarbital. Clin Pediatr (Phila)1999;38:387-394.
  20. Chang YC, Guo NW, Huang CC, et al. Neurocognitive attention and behavior outcome of school-age children with a history of febrile convulsions: a population study. Epilepsia2000;41: 412-420.
  21. Levin W, Kuntzman R, Conney AH. Stimulatory effect of phenobarbital on the metabolism of the oral contraceptive 17-alpha-ethynylestradiol-3-methyl ether (Mestranol) by rat liver microsomes. Pharmacology1979;19:255-294.
  22. Toone BK, Wheeler M, Fenwick PBC. Sex hormone changes in male epileptics. Clin Endocrinol1980;12:391-395.
  23. Taylor DC. Sexual behavior and temporal lobe epilepsy. Arch Neurol1969;21:510-516.
  24. Svensmark O, Buchthal F. Accumulation of phenobarbital in man. Epilepsia1963;4:199-206.
  25. Hollister LE. Nervous system reactions to drugs. Ann NY Acad Sci1965;123:342-353.
  26. Isbell H, Fraser HF. Addiction to analgesics and barbiturates. Pharmacol Rev1959;2:355-397.
  27. Chadwick D. Does withdrawal of different antiepileptic drugs have different effects on seizure recurrence? Further results from the MRC Antiepileptic Drug Withdrawal Study.Brain1999; 122:441-448.
  28. Desmond MM, Schwanecke RP, Wilson G, et al. Maternal barbiturate utilization and neonatal withdrawal symptomatology. J Pediatr1972;80:190-197.
  29. Bleyer WA, Marshall RE. Barbiturate withdrawal syndrome in a passively addicted infant. JAMA1972;221:185-186.
  30. Kuhnz W, Koch H, Helge H, et al. Primidone and phenobarbital during lactation period in epileptic women: total and free drug serum levels in the nursed infants and their effects on neonatal behavior. Dev Pharmacol Ther1988;11: 147-154.
  31. Melchior JC, Svensmark O, Trolle D. Placental transfer of phenobarbitone in epileptic women, and elimination in newborns. Lancet1967;2:860-861.
  32. Hawkins CF, Meynell MJ. Macrocytosis and macrocytic anemia caused by anticonvulsant drugs. Am J Med1958;27:45-63.
  33. Reynolds EH. Chronic antiepileptic toxicity: a review. Epilepsia1974;16:319-352.
  34. Reynolds EH, Travers RD. Serum anticonvulsant concentrations in epileptic patients with mental symptoms: a preliminary report. Br J Psychiatry1974;124:440-445.
  35. Jensen ON, Olesen OV. Subnormal serum folate due to anticonvulsive therapy. Arch Neurol1970;22:181-182.
  36. Mattson RH, Gallagher BB, Reynolds EH, et al. Folate therapy in epilepsy: a controlled study. Arch Neurol1973;29:78-81.
  37. Reynolds EH, Mattson RH, Gallagher BB. Relationships between serum and cerebrospinal fluid anticonvulsant drug and folic acid concentrations in epileptic patients.Neurology1972; 22:841-844.
  38. Klipstein FA. Subnormal serum folate and macrocytosis associated with anticonvulsant drug therapy. Blood1964;23: 68-86.
  39. Yerby MS, Leavitt A, Erikson DM, et al. Antiepileptics and the development of congenital anomalies. Neurology1992;42[Suppl 5]:132-140.
  40. Mountain KR, Hirsh J, Gallus AS. Neonatal coagulation defect due to anticonvulsant drug treatment in pregnancy. Lancet1970;1:265-268.
  41. Renzulli P, Tuchschmid P, Eich G, et al. Early vitamin K deficiency bleeding after maternal phenobarbital intake: management of massive intracranial haemorrhage by minimal surgical intervention. Eur J Pediatr1998;157:663-665.
  42. Christiansen C, Rodbro P, Lund M. Incidence of anticonvulsant osteomalacia and effect of vitamin D: controlled therapeutic trial. BMJ1973;4:695-701.
  43. Stamp TCB. Effects of long-term anticonvulsant therapy on calcium and vitamin D metabolism. Proc R Soc Med1974;67: 64-68.
  44. Chung S, Ahn C. Effects of anti-epileptic drug therapy on bone mineral density in ambulatory epileptic children. Brain Dev1994;16:382-385.
  45. Richens A, Rowe DJF. Disturbance of calcium metabolism by anticonvulsant drugs. BMJ1970;4:73-76.
  46. Offermann G, Pinto V, Kruse R. Antiepileptic drugs and vitamin D supplementation. Epilepsia1979;20:3-15.



  1. Muuronen A, Kaste M, Nikkila EA, et al. Mortality from ischaemic heart disease among patients using anticonvulsive drugs: a case-control study. BMJ1985;291:1481-1483.
  2. Luoma PV. Microsomal enzyme induction, lipoproteins and atherosclerosis. Pharmacol Toxicol1988;62:243-249.
  3. Luoma PV. Gene activation, apolipoprotein A-1/high density lipoprotein, atherosclerosis prevention and longevity. Pharmacol Toxicol1997;81:57-64.
  4. Schwaninger M, Ringleb P, Annecke A, et al. Elevated plasma concentration of lipoprotein (a) in medicated epileptic patients. J Neurol2000;247:687-690.
  5. Aynaci FM, Orhan F, Orem A, et al. Effect of antiepileptic drugs on plasma lipoprotein (a) and other lipid levels in childhood. J Child Neurol2001;16:367-369.
  6. Eiris J, Novo-Rodriguez MI, Del Rio M, et al. The effects on lipid and apolipoprotein serum levels of long-term carbamazepine, valproic acid and phenobarbital therapy in children with epilepsy. Epilepsy Res2000;41:1-7.
  7. Schwaninger M, Ringleb P, Winter R, et al. Elevated plasma concentrations of homocysteine in antiepileptic drug treatment. Epilepsia1999;40:345-350.
  8. Battino D, Dukes G, Perucca E. Anticonvulsants. In: Dukes MNG, Aronson JK, eds. Meyler's side effects of drugs.Amsterdam: Elsevier Science BV, 2000:166-197.
  9. Falasca GF, Toly TM, Reginato AJ, et al. Reflex sympathetic dystrophy associated with antiepileptic drugs. Epilepsia1996; 35:396-399.
  10. Maillard G, Renard G. Un nouveau traitement de l'epilepsie: la phenylethylmalonyturee. Presse Med1925;33:315-317.
  11. Beriel L, Barbier J. Le rhumatisme gardenalique. Presse Med1934;42:67-69.
  12. Mattson RH, Cramer JA, McCutchen CB, and the VA Epilepsy Cooperative Study Group. Barbiturate related connective tissue disorders. Arch Intern Med1973;149:911-914.
  13. Critchley EMR, Vakil SD, Hayward HW, et al. Dupuytren's disease in epilepsy: result of prolonged administration of anticonvulsants. J Neurol Neurosurg Psychiatry1976;39: 498-503.
  14. Janz D, Piltz U. Frozen shoulder induced by primidone. In: Oxley J, ed. Antiepileptic therapy: chronic toxicity of antiepileptic drugs.New York: Raven Press, 1983:155-159.
  15. Froscher W, Hoffman F. Dupuytren's contracture in patients with epilepsy: follow-up study. In: Oxley J, ed. Antiepileptic therapy: chronic toxicity of antiepileptic drugs.New York: Raven Press, 1983:147-154.
  16. Noble J. Connective tissue disorders: discussion. In: Oxley J, ed. Antiepileptic therapy: chronic toxicity of antiepileptic drugs.New York: Raven Press, 1983:169-173.
  17. Horton P, Gerster JC. Reflex sympathetic dystrophy syndrome and barbiturates: a study of 25 cases treated with barbiturates compared with 124 cases treated without barbiturates. Clin Rheumatol1984;3:493-500.
  18. Zimmerman HJ. Drug-induced liver disease. Drugs1978;16: 25-45.
  19. Richens A. The clinical consequences of chronic hepatic enzyme induction by anticonvulsant drugs. Br J Clin Pharmacol1974;1:185-187.
  20. Baillie TA. Metabolic activation of valproic acid and drug-mediated hepatotoxicity: role of the terminal olefin, 2-n-propyl-4-pentenoic acid. Chem Res Toxicol1988;1:195-199.
  21. Perucca E. Clinical implications of hepatic microsomal enzyme induction by antiepileptic drugs. Pharmacol Ther1987;33: 139-144.
  22. Burstein S, Klaiber E. Phenobarbital induced increase in 6-beta-hydroxycortisol excretion: clue to its significance in human urine. J Clin Endocrinol Metab1965;25:293-296.
  23. Brooks SM, Werk EE, Ackerman SJ, et al. Adverse effects of phenobarbital on corticosteroid metabolism in patients with bronchial asthma. N Engl J Med1972;286:1125-1128.
  24. Janz D, Schmidt D. Antiepileptic drugs and failure of oral contraceptives. Lancet1974; 1:1113.
  25. MacDonald MG, Robinson DS. Clinical observations of possible barbiturate interference with anticoagulation. JAMA1968;204:95-100.
  26. Granick S. Hepatic porphyria and drug-induced or chemical porphyria. Ann NY Acad Sci1965;123:188-197.
  27. Yeung CY, Tam LS, Chan A, et al. Phenobarbitone prophylaxis for neonatal hyperbilirubinemia poisoning. Pediatrics1971;48: 372-376.
  28. Cramer JA, Scheyer RD. The effect of alcohol use on antiepileptic drugs. In: Porter RJ, Mattson RH, Cramer JA, et al., eds. Alcohol and seizures.Philadelphia: FA Davis, 1990:241-250.
  29. Conney AH, Jacobson M, Schneidman K, et al. Induction of liver microsomal cortisol 6β-hydroxylase by diphenylhydantoin or phenobarbital: an explanation for the increased excretion of 6-hydroxycortisol in humans treated with these drugs. Life Sci1965;4:1091-1098.
  30. Rubin E, Hutterer F, Lieber CS. Ethanol increases hepatic smooth endoplasmic reticulum and drug-metabolizing enzymes. Science1968;159:1469-1470.
  31. Rubin E, Lieber CS. Hepatic microsomal enzymes in man and rat: induction and inhibition by ethanol. Science1968;162: 690-691.
  32. Cramer JA, Cohn G, Meggs L. Effect of phenobarbital and heroin metabolism in the rat. Fed Proc1975;34:814.
  33. Hollister LE. Nervous system reactions to drugs. Ann NY Acad Sci1965;123:342-353.
  34. Clemmensen J, Fuglsang-Frederiksen V, Plum CM. Are anticonvulsants oncogenic? Lancet1974;1:705-707.
  35. White SJ, McLean AEM, Howland C. Anticonvulsant drugs and cancer: a cohort study in patients with severe epilepsy. Lancet1979;2:458-461.
  36. Olsen HH, Boice JD, Jensen JP, et al. Cancer among epileptic patients exposed to anticonvulsant drugs. J Natl Cancer Inst1989;81:803-808.
  37. Schmidt RP, Wilder BJ. Epilepsy.Philadelphia: FA Davis. 1968.
  38. Welton DG. Exfoliative dermatitis and hepatitis due to phenobarbital. JAMA1950;143:232-234.
  39. Schlienger RG, Shear NH. Antiepileptic drug hypersensitivity syndrome. Epilepsia1998;39[Suppl 7]:S3-S7.
  40. McGeachy TE, Bleemer WE. The phenobarbital sensitivity syndrome. Am J Med1953;14:600-604.
  41. Yatzidis H. The use of ion exchange resins and charcoal in acute barbiturate poisoning. In: Matthew H, ed. Acute barbiturate poisoning. Amsterdam: Excerpta Medica, 1971:223-232.
  42. Stuttgen G. Toxic epidermal necrolysis provoked by barbiturates. Br J Dermatol1973;88:291-293.
  43. Ruble J, Matsuo F. Anticonvulsant-induced cutaneous reactions. Incidence, mechanisms and management. CNS Drugs1999;12:215-236.
  44. Alarcon-Segovia D. Drug-induced lupus syndromes. Mayo Clin Proc1966;44:664-681.
  45. Schain RJ, Watanabe K. Origin of brain growth retardation in young rats treated with phenobarbital. Exp Neurol1976;50: 806-809.
  46. Hiilesmaa VK, Teramo K, Granstrom ML, et al. Fetal head growth retardation associated with maternal antiepileptic drugs. Lancet1981;2:165-167.



  1. Bergey GK, Swaiman KF, Schreir BK, et al. Adverse effects of phenobarbital on morphological and biochemical development of fetal mouse spinal cord neurons in culture.Ann Neurol1981; 9:584-589.
  2. Neale EA, Sher PK, Graubard BI, et al. Differential toxicity of chronic exposure to phenytoin, phenobarbital, or carbamazepine in cerebral cortical cell cultures. Pediatr Neurol1985;1:143-150.
  3. Serrano EE, Kunis DM, Ransom BR. Effects of chronic phenobarbital exposure on cultured mouse spinal cord neurons. Ann Neurol1988;24:429-438.
  4. Koch S, Titze K, Zimmermann RB, et al. Long-term neuropsychological consequences of maternal epilepsy and anticonvulsant treatment during pregnancy for school-age children and adolescents. Epilepsia1999;40:1237-1243.
  5. Reinisch JM, Sanders SA, Mortensen EL, et al. In utero exposure to phenobarbital and intelligence deficits in adult men. JAMA1995;274:1518-1825.
  6. Chatot CL, Klein NW, Clapper ML, et al. Human serum teratogenicity studied by rat embryo culture: epilepsy, anticonvulsant drugs, and nutrition. Epilepsia1984;25:205-216.
  7. Speidel BD, Meadow SR. Maternal epilepsy and abnormalities of the fetus and newborn. Lancet1972;2:839-843.
  8. Nelson MM, Forfar JO. Associations between drugs administered during pregnancy and congenital abnormalities of the fetus. BMJ1971;1:523-527.
  9. Fedrick J. A report from the Oxford record linkage study. BMJ1973;1:442-448.
  10. Shapiro S, Hartz SC, Siskind V, et al. Anticonvulsants and parental epilepsy in the development of birth defects. Lancet1976;1:272-275.
  11. Mastroiacovo P, et al. Fetal growth in the offspring of epileptic women: results of an Italian multicentric cohort study. Acta Neurol Scand1988;78:110-114.
  12. Dansky L, Andermann NC, Sherwin A, et al. Maternal epilepsy and congenital malformation: a prospective study with monitoring of plasma anticonvulsant levels during pregnancy. Neurology1980;30:438.
  13. Dravet C, Julian C, Legras C, et al. Epilepsy, antiepileptic drugs, and malformations in children of women with epilepsy: a French prospective cohort study. Neurology1992;42[Suppl 5]: 75-82.
  14. Samren EB, van Duijn CM, Christiaens GC, et al. Antiepileptic drug regimens and major congenital abnormalities in the offspring. Ann Neurol1999;46:739-746.
  15. Waters CH, Belai Y, Gott PS, et al. Outcomes of pregnancy associated with antiepileptic drugs. Arch Neurol1994;51:250-253.
  16. Kaneko S, Battino D, Andermann E, et al. Congenital malformations due to antiepileptic drugs. Epilepsy Res1999;33:145-158.
  17. Berman LB, Jeghers HJ, Schreiner GE, et al. Hemodialysis, an effective therapy for acute barbiturate poisoning. JAMA1956;161:820-827.
  18. Mawer GE, Lee HA. Value of forced diuresis in acute barbiturate poisoning. BMJ1968;2:790-792.
  19. Goldberg MJ, Berlinger WG. Treatment of phenobarbital overdose with activated charcoal. JAMA1982;247:2600-2601.