Drug Safety 1996 Dec; 15 (6): 378-393

DRUG EXPERIENCE

011 4-5916/96/OO12-o378/S08 .001O

© Adis International Limited. All rights reserved .

A Comparative Review of the Adverse Effects of Anticonvulsants in Children with Epilepsy Sheila J. Wallace Department of Paediatric Neurology, University Hospital of Wales, Cardiff, Wales

Contents Summary ..................... . 1. Basic Mechanisms of Action of Anticonvulsants 2. Review of Organ Systems Involved .. 3. Special Considerations for Neonates . 4. Systematic Review of Adverse Effects . 4.1 Haematological Adverse Effects 4.2 Endocrine Adverse Effects . . . 4.3 Rash and Other Skin Disorders 4.4 Systemic Adverse Effects ... 4.5 Metabolic Disturbances . . . . 4.6 Gastroenterological Adverse Effects . 4.7 Renal Adverse Effects . . . . . . . . . 4.8 Cosmetic Adverse Effects . . . . . . . 4.9 Adverse Effects on the Nervous System 5. Exacerbation of Seizures 6. Conclusion . . . . . . . . . . . . . . . . . . .

Summary

378 379 381 381 381 381 382 383 383 384

385 387 387 387 390 391

Phenobarbital (phenobarbitone) and phenytoin are the most useful anticonvulsants in neonates because adverse effects are most readily reversed when these drugs are used. Most anticonvulsants are very rarely associated with haematological adverse effects. Platelet function is particularly vulnerable to valproic acid (sodium valproate) therapy. Barbiturates and phenytoin can precipitate metabolic bone disease. Although very infrequent, lymphadenopathy is most common with phenytoin, and lupuslike illnesses with ethosuximide. Valproic acid may precipitate underlying metabolic disorders. Nephrolithiasis can occur with topiramate. Liver disease is most likely with felbamate or valproic acid, but can occur with other anticonvulsants. Valproic acid and ethosuximide are the main precipitants of gastrointestinal symptomatology; while valproic acid and vigabatrin are frequently associated with excessive bodyweight gain. Rash is most likely to occur with barbiturates, but there is a high risk of this adverse effect if large doses of lamotrigine are given with valproic acid. Adverse

379

Anticonvulsants in Childhood Epilepsy

cosmetic effects are most likely with phenytoin, but valproic acid may cause alopecia. All anticonvulsants may cause unwanted neurological effects: when they occur, diplopia is usually precipitated by carbamazepine; tremor by valproic acid; and other motor disturbances are probably most common with phenytoin. Most anticonvulsants can cause drowsiness. Phenobarbital leads anticonvulsants as a cause of behavioural difficulties. Effects of anticonvulsants on cognitive function are difficult to assess, but subtle changes have been reported for all anticonvulsants in use up to the 1980s. Compared with other anticonvulsant drugs, phenytoin and felbamate are more often discontinued as a result of unwanted effects. A comparative study of the adverse effects of anticonvulsants, used in childhood epilepsies, poses some difficulties. In well conducted, closelymonitored trials, all adverse events outside totally normal health are recorded, so that common childhood symptomatologies tend to be noted as adverse events. In the absence of a placebo group - regrettably, the usual state of affairs when a trial of a new anticonvulsant is being conducted in children - the background incidence of such adverse events is not identified, and their significance in relation to the anticonvulsant is difficult to assess. Many children who are exposed to anticonvulsants are too young or developmentally delayed to make specific complaints in respect of effects that they find uncomfortable or irritating. Some parameters of health, such as liver function tests (especially alkaline phosphatase), may be less stable in childhood than in adults, leading to difficulties in interpretation of the significance of any change in liver function values. Behavioural difficulties seen in children treated with anticonvulsants can be either be primary, or secondary to feelings of dizziness, headache, or general debility etc, or abdominal discomfort, which cannot be clearly defined. In children, and particularly in infants, the brain is a rapidly developing organ. Tests of cognitive ability, which may be appropriate at an early age, are unlikely to be suitable for older children, so it is often not possible to extrapolate from one series of assessments to another. When serious adverse events occur, it is usual for these to appear as single or multiple case reports, and the actual incidence of such episodes is © Adis International Limited. All rights reserved.

difficult to estimate. Nevertheless, it can be inferred that, if such serious events present in association with a much used anticonvulsant, and are considered worthy of publication, they must be very rare. In an attempt to draw some conclusions about the relative risks of the various anticonvulsants in common use, and those that have been recently introduced, adverse effects are discussed according to the organ system involved. A statement about the relative risks of abnormalities in the organ system under review is given at the end of each section. For the purposes of this review, only those adverse effects that occur at conventional (rather than toxic) dosages will be considered - i.e. those events that occur during 'optimal' management.

1. Basic Mechanisms of Action of Anticonvulsants The 3 main possible mechanisms Df action of anticonvulsants that have been identified are: • an increase in y-aminobutyric acid (GAB A)ergic inhibition at the GABA A receptor complex; • blockade of sodium channels in their inactivated form leading to 'use-dependent' suppression of receptive discharges; • inhibition of calcium T channels. Phenobarbital (phenobarbitone), benzodiazepines, vigabatrin and tiagabine work primarily through GAB A-related mechanisms. Primidone, by virtue of its breakdown to phenobarbi tal (as well as phenylethylmalondiamide) must also work via GABA inhibition. There is some evidence that bromides and topiramate also have GAB A-related Drug Safety 1996 Dec; 15 (6)

380

inhibitory properties. Phenytoin, carbamazepine and lamotrigine act via 'use-dependent' blockade of voltage-gated sodium channels. As well as its GAB A-related effects, topiramate also has effects on sodium channels. Ethosuximide inhibits calcium T channels. The precise mechanisms of action of valproic acid (sodium valproate), gabapentin and felbamate have not been identified, but both gabapentin and felbamate are known to have ac~ tions at both GAB A receptor and sodium channel sites. Prediction of adverse effects on the basis of mechanisms of action of anticonvulsants is difficult, particularly for effects outside the central nervous system. Broadly speaking, those anticonvulsants that act primarily via GABA-related mechanisms are more likely to precipitate unacceptable behavioural difficulties than those that act mainly on sodium or calcium channels. On the other hand, rash is a possible adverse effect with all drugs that have primary action via sodium channels, although rash is not confined to this group (for example, rash also occurs with barbiturates). Some systemic effects can be predicted on the basis of the presence or absence of induction of enzymes, particularly in the liver, but sometimes elsewhere. Cytochrome P450 isoenzymes are most commonly affected and such induction is dosedependent. The main enzyme inducers are phenobarbital, primidone, carbamazepine, and phenytoin. Although the results of enzyme induction are similar with these anticonvulsants, the isoenzymes induced by carbamazepine and phenytoin are not the same as those induced by barbituratesJ'l Valproic acid, ethosuximide, lamotrigine, gabapentin, and vigabatrin are not enzyme inducers. Enzyme inducers can increase the elimination of concurrently administered drugs; particularly other anticonvulsants (such as lamotrigine), corticosteroids, anticoagulants and some antibacterials. Clearly, such induction can be of clinical importance, but tends to be of less relevance in childhood than in older age groups. Induction of the metabolism of endogenous substrates can cause problems. An example is increased © Adis International Limited. All rights reserved.

Wallace

metabolism of cholecalciferol (vitamin 03), leading to anticonvulsant-induced rickets. Enzyme induction may also contribute to the following anticonvulsant-associated disorders: folic acid deficiency; vitamin K-responsive haemorrhagic disorder in infants of mothers with epilepsy; altered metabolism of steroid and thyroid hormones; increased plasma levels of (X,-acid glycoprotein, sex hormone-binding globulin, y-glutamyl transferase and alkaline phosphatase; precipitation of acute episodes in patients with acute intermittent porphyria; and enhancement of insulin-mediated glucose disposal ratePl Enzyme inhibition can lead to interference with the metabolism of other drugs. Valproic acid is the chief example of an enzyme inhibitor amongst anticonvulsants, but other anticonvulsant concentrations and activities can be affected by concurrent use of drugs from other classes for a co-existing illness. The result may be unexpected anticonvulsant toxicity at relatively low dosages. Phenytoin, because of its saturation kinetics, is most vulnerable to enzyme inhibitors.l21 In theory, the use of substances that alter neurotransmission might interfere with normal brain development at critical periods in early childhood. Synaptogenesis is very active in the first year of life. Initially, there is an 'overproduction' with 30 to 40% more synapses in the brain in the second and third years of life than later. Pruning then occurs and by 5 years of age, synaptic counts are similar to those in adults. There is no evidence that anticonvulsants available before the 1980s have adverse effects on brain development, but more recently developed anticonvulsants are more specifically targeted and thus may be more likely to affect specific chemical aspects in brain maturation. The available evidence - for example, the use of vigabatrin for infantile spasms - suggests that control of seizures far outweighs any possible adverse effects on neurotransmission. However, very careful monitoring of infants treated with newer or future anticonvulsants is clearly necessary. Drug Safety 1996 Dec; 15 (6)

Anticonvulsants in Childhood Epilepsy

2. Review of Organ Systems Involved Before exposure to humans, all anticonvulsants will have been tested in animals, and, before use in children, adults will have taken part in clinical trials. Thus, children are in a somewhat privileged position. Nevertheless, immaturity of many organ systems, and the possibility of precipitating otherwise subclinical inborn errors of metabolism, make children more vulnerable in areas that are less likely to be a problem in older individuals. It is particularly important to watch for multisystem disease, and haematological, renal and hepatic complications. Endocrine effects could be relevant to growth and pubertal changes. Dermatological changes, especially rashes, can be but one indication of a more generalised reaction. Cosmetic problems, although important at any age, are of particular concern in adolescence. Gastrointestinal disturbances may lead to reduced appetite, with secondary effects on growth or bone metabolism. Disturbance of neurological competence may make children less competitive than their peers, more dependent, and cause difficulties with skilled movements such as feeding, dressing and writing. Alterations in behaviour can increase family tensions, and make arrangements for school placement and childminding more difficult. Studies of cognition are often difficult to interpret, but are very important in the context of children with epilepsy, who frequently do less well than expected in school.

3. Special Considerations for Neonates The drug that is usually chosen for the initial treatment of neonatal seizures is phenobarbital,13] but phenytoin may also be used. Both of these drugs can increase bleeding tendencies by interfering with vitamin K-dependent clotting factors,14.5J but this problem is treatable with phytomenadione (vitamin K I ) administration. Although prenatal phenytoin may produce migrational disturbances in the cerebellum, short term treatment with pheny© Adis International Limited. All rights reserved.

381

toin in neonates does not seem to cause cerebellar abnormalities. Many preparations of diazepam, another possible choice of drug for treating neonatal seizures, contain sodium benzoate, which could increase the plasma level of unconjugated bilirubin,16] but this does not seem to be a problem in practice. In addition, as a result of very slow metabolism, the elimination half-life of diazepam in neonates is several days. Lorazepam has been successfully used to treat refractory neonatal seizures, without apparent unwanted effects.[7·8] Paraldehyde, which has also been used successfully in neonatal seizures, is a caustic material. Its excretion through the lungs may cause damage to an already compromised respiratory system and it should be avoided if possible. 19 ] Neonatal seizures are particularly likely to be secondary to metabolic upsets. On theoretical grounds, the use of valproic acid in the neonatal period is probably unwise, since disorders of the urea cycle and fatty acid metabolism may be precipitated by this anticonvulsant. IIO ) In summary, possible adverse effects of anticonvulsants used in the neonatal period are most readily reversible when either phenobarbital or phenytoin is used.

4. Systematic Review of Adverse Effects 4.1 Haematological Adverse Effects

On the whole, adverse haematological effects are more commonly reported in association with long-established, rather than recently introduced, anticonvulsants. The experience with felbamatel II] is the exception. Clearly, long-established drugs have been used in many more children, and reporting may, to some extent, reflect patient exposure. Nevertheless, certain types of haematological disorder seem to be more likely with some anticonvulsants rather than others. Aplastic anaemia has been reported to occur at an annual rate of 1 in 2000 patients treated with felbamate. [tl] It is described as a very rare occurrence with carbamazepine,112] phenytoin l4 ] and Drug Safety 1996 Dec; 15 (6)

382

Wallace

ethosuximide.!13] Megaloblastic anaemia secondary to folic acid deficiency may complicate the use of phenobarbital,15] primidone l14 ] or phenytoin.!4] This adverse effect occurs in less than 1% of patients treated with phenobarbital,15] and is much rarer than mild macrocytosis, which can be seen in 50% of patients of all ages treated with phenytoin.!4] However, it is possible that children are less likely to develop macrocytosis than adults; in 1 study none of the 55 children receiving phenytoin and only 4% of those receiving phenobarbital experienced macrocytosis. I 15] Mattson[l6] has stated that leucopenia 'can occur with any antiepileptic drug' and that it is no more common with any single agent; however, he was referring mainly to phenobarbital, carbamazepine, phenytoin and valproic acid. Nevertheless, cyclical neutropenia, of doubtful clinical significance, is more likely to be associated with carbamazepine than with other anticonvulsants. In one study,11 7] 35% of children had a reduced white blood cell count (2000 to 4000/mm 3), but in half of them, the finding was transient, and in none did it appear to be clinically significant. Persistent leucopenia was found in only I % of those children taking carbamazepine in another study'! 15] Agranulocytosis may very rarely occur with phenytoin,14] carbamazepine l12 ] or ethosuximide.!13] Eosinophilia can be found as part of a systemic reaction to phenytoin, I 18] or as an otherwise isolated change in 20% of children receiving carbamazepine.112 ]

Thrombocytopenia is a very rare complication of phenytoin[4] or carbamazepine l12 ] use. Valproic acid is more likely to specifically affect platelet numbers and function .[10,19,20] A combination of inhibition of platelet aggregation and fibrinogen depletion can lead to an increased tendency to bleed during surgery in children receiving valproic acid. llO ] Alterations in platelet function appear to be dose-related for this agent.! 19,20,2 I] Haematological adverse effects do not seem to occur with vigabatrin,122] lamotrigine,123] gabapentin,[24] tiagabine l25 ] or topiramate.126] In summary, felbamate seems most likely to cause very serious haematological adverse effects. Note should be taken of abnormalities of platelet function with high-dose valproic acid. Otherwise, the haematological system is adversely affected by anticonvulsants on only very rare occasions. For an overview, adverse haematological effects are summarised in table I. 4.2 Endocrine Adverse Effects

Various alterations in plasma levels of the products of endocrine glands have been reported in association with anticonvulsant treatment. 14 ,12] However, the finding that vertical growth is entirely normal in childhood despite anticonvulsant therapyl27] strongly suggests that such changes are not of clinical importance. Water retention secondary to the antidiuretic effects of carbamazepine, and associated with a reduction in plasma sodium levels,112] is only rarely

Table I. Incidence of haematological adverse effects associated with anticonvulsant drugs PBT

Adverse Effect

PRI

PHE

VAL

VR

Aplastic anaemia Megaloblastic anaemia

R

Mild macrocytosis

+

eBZ

ETH

FBM

VR

VR

R

VR

R

R

±

Leucopenia (cyclical) Agranulocytosis

VR

+ VR

Eosinophilia

VR

+++

Thrombocytopenia

VR

Abnormal platelet function

R

VR

D

=

= = =

= =

=

=

carbamazepine; D dose-related; ETH ethosuximide; FBM felbamate; PBT phenobarbital Abbreviations and symbols: eBZ (phenobarbitone); PHE phenytoin; PRI = primidone; R rare «0.001%); VAL valproic acid (sodium valproate) ; VR very rare (single case reports); ± = 0.001 to 0.99%; + = 1.00 to 4.99%; +++ >10%.

=

© Adis International Umited. All rights reserved.

=

Drug Safety 1996 Dec : 15 (6)

383

Anticonvulsants in Childhood Epilepsy

Table II. Incidence of endocrine adverse effects associated with anticonvulsant drugs PBT

Adverse effect

PRI

PHE

VAL

Antidiuretic effects Metabolic bone disease

R

R

Secondary amenorrhoea

CBl

OXCBl

±

+

R VR

Abbreviations and symbols: CBl =carbamazepine; OXCBl =oxcarbazepine; PBT =phenobarbital (phenobarbitone); PHE =phenytoin; PRI

=

primidone; R to 4.99%.

=rare «0.001%); VAL =valproic acid (sodium valproate); VR =very rare (single case reports); ± =0.001 to 0.99%; + =1.00

a problem. Oxcarbazepine has a more potent antidiuretic effect than carbamazepine but this seems to be more of a problem in the elderly than in childhood,l28] Metabolic bone disease may be precipitated in children on poor diets who receive long term phenobarbital,[16] primidone,[14] or phenytoin therapy.[4] Valproic acid is rarely associated with secondary amenorrhoea. [10] The endocrine changes associated with anticonvulsant therapy are summarised in table II.

4.3 Rash and Other Skin Disorders Skin rashes are amongst the most common adverse effects precipitated by anticonvulsants. They have been reported with phenytoin,[16] phenobarbitaJ,!16] primidone,l14] carbamazepine,l16] ethosuximide,[13]lamotrigine,[23,29] felbamate[30,31] and bromides,[32] and very rarely with valproic acid,l16] Neither gabapentin nor vigabatrin produce rashes. Oxcarbazepine is less likely to cause a rash than carbamazepine, but about 25% of those children sensitive to carbamazepine will also react to oxcarbazepine. [28] In almost all affected patients, the rash is maculopapular, although erythema multiforme has been reported with ethosuximide)13] StevensJohnson syndrome may be a rare adverse effect in patients treated with any of the above anticonvulsants. [13,14.16.18.23,29-31 ,33 1 It is not always easy to estimate the actual incidence of rash from reports in the literature, and figures often vary widely. In a comparative study,[15] the reported incidence of rash was 4% with phenytoin, 5% with carbamazepine, 10% with phenobarbital and less than I % with valproic acid; no rash was recorded in the small numbers of chil© Adis International Limited. All rights reserved.

dren treated with ethosuximide or benzodiazepines. Another childhood study[17] suggested that rash occurred in 2% of children treated with carbamazepine. The propensity to develop rash with lamotrigine has been studied in most detail. There is no doubt that combined treatment with lamotrigine and valproic acid substantially increases the risk of rash. [23,29] In an overview of patients of all ages,[23] 2.5% of patients receiving valproic acid and 1% of those not receiving valproic acid developed rashes when lamotrigine was added. The rate was much higher in a childhood study, [29] in which, overall, 25% of children given lamotrigine developed rashes, but only 2 (2.5%) of the 79 children not taking concurrent valproic acid were affected. The recognition that very low dosages and slow dosage escalation of lamotrigine can avert most skin reactions has led to a much lower frequency of rash in more recent studies. Although rash was reported to be the most frequent reason for discontinuation of felbamate in studies involving children,!311 a rate of 6%, compared with 7% with placebo, has been recorded.[30] The rates of rash associated with anticonvulsants are summarised in table III.

4.4 Systemic Adverse Effects Generalised reactions, including systemic lupus erythematosus, have been reported in association with phenytoin,[15.18,34] carbamazepine,[12,15] ethosuximide[13] and lamotrigine,!35] Only one case of systemic lupus erythematosus secondary to primidone has been authenticated. [14] Fever and/or pseudolymphoma were reported in 2% of children treated with phenytoin.!15] Using molecular biological Drug Safety 1996 Dec; 15 (6)

384

Wallace

Table III. Incidence of dermatological adverse effects associated with anticonvulsant drugs Adverse effect

PBT

PAl

PHE

CBZ

+++

++

+

++

ETH

LTG

FBM

VAL

++

++

VA

A

A

Bromides

OXCBZ

Aash maculopapular acneiform Stevens-Johnson syndrome

+

±

erythema multiforme + VA

VA

VA

+D VA

A

Abbreviations and symbols: CBZ =carbamazepine; D =dose-related; ETH =ethosuximide; FBM =felbamate; LTG =lamotrigine; OXCBZ =oxcarbazepine; PBT =phenobarbital (phenobarbitone); PHE =phenytoin; PAl =primidone; A =rare «0.001%); VAL =valproic acid (sodium valproate); VA =very rare (single case reports); ± =0.001 to 0.99%; + =1.00 to 4.99%; ++ =5.00 to 9.99%; +++ => 10%.

techniques, pseudolymphomata seen with phenytoin have been demonstrated to be hyperplastic, rather than malignant. (34 ) Reports of patients with vasculitis, myocarditis, membranous glomerulopathy, pseudolymphoma, tubulo-interstitial nephritis, and systemic lupus erythematosus have been covered in an overall review of carbamazepine toxicity.(12) In addition, carbamazepine can have specific effects on cardiac conduction: 112 ) older patients with sick sinus syndrome are most at risk, but it is suggested that children with tuberous sclerosis and cardiac rhabdomyomata could be affected. Three types of lupus-like reactions have been reported in association with ethosuximide:I)3) the presence of antinuclear antibodies only; classical lupus-like illness; and clinical and chemical evidence of immunological disorder, including the nephrotic syndrome, but with negative results in blood tests for antinuclear antibody and lupus preparations. With lamotrigine, cases of disseminated intravascular coagulation, rhabdomyolysis, renal failure and rash have occurred extremely rarely.(35) In general, systemic reactions are very rare, regardless of the anticonvulsant used. Their frequencies are shown in table IV. 4.5 Metabolic Disturbances

Disturbances in metabolism and clinical precipitation of underlying inborn metabolic errors are generally confined to patients treated with valproic acid and are summarised in table V. Specific effects of valproic acid on the liver are considered in section 4.6.1. © Adis International Limited. All rights reserved.

4.5.1 Valproie Aeld (Sodium Valproote) and Metabolites

Hyperammonaemia without other evidence of disease may be found in 1% of children taking valproic acid.(15 ) However, in both partial ornithine transcarbamylase deficiency, and in carriers of this condition, clinically significant hyperammonaemia can be precipitated by valproic acid therapy. I) OJ The metabolites of valproic acid seem to be as important in determining its adverse effect profile as they are in contributing to its anticonvulsant properties.(36) These metabolites are similar in structure to endogenous organic acids, and some have been shown to be pharmacologically active, whereas others appear to be teratogenic or hepatotoxic. Valproic acid is involved in a number of conjugation reactions. Abnormally high concentrations

Table IV. Incidence of systemic adverse reactions associated with anticonvulsant drugs Adverse reaction

PAl

Systemic lupus erythematosus

VA

Pseudolymphoma

PHE

CBZ

ETH

VA

A

LTG

VA

+

Vasculitis

VA

Myocarditis

VA

Renal disease

VR

Adverse effects on cardiac conduction

VR

VR

Disseminated intravascular coagulation

VR

=

= =

Abbreviations and symbol: CBZ carbamazepine; ETH ethosuximide; LTG lamotrigine; PHE phenytoin; PRI primidone; R rare «0.001 %); VR very rare (single case reports); + 1.00 to 4.99%.

=

=

=

=

=

Drug Safety 1996 Dec; 15(6)

Anticonvulsants in Childhood Epilepsy

of valproic acid conjugates, mainly rearranged glucuronide isomers, have been reported in a child with associated hepatobiliary and renal problems. Secondary carnitine deficiency[37] and, very rarely, Reye-like syndromes, which have been observed in valproic acid-treated children, could be precipitated, or exacerbated, by diversion of carnitine to valproic acid conjugated formation. Co-administration of calcium pantothenate, acetylcysteine and levocarnitine might prevent these effects,[38] although there is no evidence in humans that they do so. The importance of conjugation of valproic acid metabolites with glycine is unclear, since hyperglycinaemia and hyperglycinuria are not associated with increased glycine levels in the cerebrospinal fluid and no clinical symptoms have been identified. Conjugation with coenzyme A (CoA) is potentially of more importance, since valproic acid-CoA inhibits several intermediate metabolic processes, and is a poor substrate for medium chain acyl-CoA hydrolases. Thus, acute illness may be precipitated in patients with medium chain acylCoA dehydrogenase deficiency.[IO] Valproic acid and endogenous lipids compete for the enzymes of mitochondrial ~-oxidation, with resultant increases in the urinary excretion of propionic acid and in plasma levels of propionatederived amino acids. Reduced ~-oxidation and increased ill-oxidation of valproic acid are reported in patients with encephalopathy[39] and hepatic failure.[40] Metabolism of valproic acid via the illoxidative (cytochrome P450) enzyme systems leads to the formation of the hepatotoxic metabolite, 4-en-valproic acid,[3] but it is not clear whether reduced ~-oxidation precedes, or is merely associated with, hepatotoxicity. The possibility that valproic acid can exacerbate the tendency of several inborn errors of metabolism to cause hepatic failure must always be considered,[41] particularly in very young, developmentally delayed children, in whom poly therapy including valproic acid seems particularly hazardous. [42] Clearly, valproic acid should be avoided when an underlying disorder of mitochondrial function, © Adis International Limited. All rights reserved.

385

Table V. Incidence of metabolic disturbances associated with valproic acid (sodium valproate) Metabolic disturbance

Incidence with valproic acid

Uncomplicated hyperammonaemia

+

Clinically significant hyperammonaemia in partial defects of urea cycle enzymes

+++

Secondary carnitine deficiency

+

Hyperglycinaemia and hyperglycinuria

±

PreCipitation of fatty acid decompensation"

++

PropioniC acidemia + propionic aciduria"

+

Reduced ~-oxidation + increased (O-oxidation" + a

In susceptible individuals.

Symbols: ± =0.001 to 0.99%; + =1.00 to 4.99%; ++ =5.00 to 9.99% +++ = >10%.

fatty acid metabolism or urea cycle function is suspected. 4.5.2 Other Anticonvulsants

Acute intermittent porphyria (AlP) may be precipitated by anticonvulsants. The incidence is between 1 and 4.99%. In patients with relative porphobilinogen deaminase deficiency, hepatic ()aminolaevulinic acid synthase is induced by barbiturates, primidone, carbamazepine, phenytoin, succinimides, trimethadione and valproic acid. Although there is some doubt as to whether or not it is safe to use nitrazepam, other benzodiazepines probably do not precipitate AIP.[43] In addition, gabapentin has been used without problems.[44] AlP should be suspected, and specifically tested for, if seizures become dramatically worse following the introduction of anticonvulsants. Other acute symptoms include abdominal pain, vomiting, weakness secondary to neuropathy, hypertension and mental changes. 4.6 Gastroenterological Adverse Effects 4.6.1 Uver Function

Clinically unimportant increases in the enzymes measured when assessing liver function are found following administration of all enzyme-inducing anticonvulsants and valproic acid.[IO] Such elevations were noted in 16% of 90 children receiving carbamazepine. lI7 ] In a comparative study,ll51 raised levels of liver enzymes were found in the plasma Drug Safety 1996 Dec; 15 (6)

386

Wallace

Table VI. Incidence of hepatic adverse effects associated with anticonvulsant drugs Adverse effect

PST

PRI

PHE

CSZ

VAL

Unimportant elevation of hepatic enzyme levels

+

+

+

+

++

VR

VR

Allergic hepatitis: granulomatous reaction Acute hepatitis + hepatocellular necrosis

VR

Toxic necrosis secondary to abnormalities in mitochondrial oxidative mechanisms (in susceptible individuals) Fatal hepatic failure

FSM

VR

VR

VR

VR

R

=carbamazepine; FSM =felbamate; PST =phenobarbital (phenobarbitone); PHE =phenytoin; PRI = primidone; R =rare «0.001%); VAL =valproic acid (sodium valproate); VR =very rare (single case reports); + =1.00 to 4.99%; ++ =5.00 Abbreviations and symbols: CSZ

to 9.99%.

of 4% of children treated with phenytoin, 6% of those treated with valproic acid and I % of those treated with carbamazepine. No child taking ethosuximide or benzodiazepines had raised liver enzyme levels, nor, rather surprisingly, did those taking phenobarbital. Neither tiagabine[25] nor gabapentin[45] cause liver changes in adults, but childhood studies are not available yet. Phenytoin can cause a granulomatous reaction in the liver, in association with allergic hepatitis: the mortality rate is 25%, and resolution is slow in survivors)18] Very rare instances of 2 types of hepatic problem can occur with carbamazepine;[12] cases of a hypersensitivity-induced granulomatous hepatitis and a direct toxic effect of carbamazepine (or its metabolites), with acute hepatitis and hepatocellular necrosis, have been cited. Soon after the introduction of valproic acid, much publicity was given to possible adverse effects on the liver.[46,47] As discussed in section 4.5 .1, some valproic acid metabolites, particularly 4-en-valproic acid, can be hepatotoxic. However, it seems probable that hepatotoxicity occurs almost exclusively in patients with underlying problems in mitochondrial oxidative metabolism. Those at greatest risk of dying of liver failure are young children with developmental delay and coincidental metabolic disorders, who receive valproic acid as polytherapy)42] It is now recognised that some deaths from liver failure, previously attributed to treatment with valproic acid, were the result of familial hepatic dysfunction. The histological changes associated with valproic acid-induced liver failure © Adls International Limited. All rights reserved.

are similar to those seen in patients with alcoholism)18] Felbamate may also cause liver failure) II] In summary, adverse events affecting the liver are rare, but are most likely to present in the very young who have predisposing conditions. Their relative frequencies are shown in table VI. 4.6.2 Gastrointestinal Adverse Effects and Pancreatitis

Minor gastrointestinal disturbances are common, and are reported with all anticonvulsants except vigabatrin. Constipation occurs in 2% of patients treated with carbamazepine,P5] which also causes diarrhoea, vomiting and nausea in a total of between 9% and 14% of children.[33,48] Higher plasma carbamazepine concentrations are associated with an increased incidence of gastrointestinal complaints)48] Rarely, carbamazepine therapy is associated with eosinophilic colitis)12] Phenobarbital and primidone may also cause problems,f I6 ,48] but valproic acid is most likely to produce digestive complaints,148] with up to 25% of patients possibly being affected. IIO ] About 25% of children treated with ethosuximide complain of nausea. fI3 ] Abdominal discomfort, anorexia and hiccups are less common. A reduction in ethosuximide dosage usually resolves these symptoms. Vomiting was reported in 1.8% of children receiving lamotrigine;129] this was most likely to occur when dosage given was high. Although anorexia is considered to be the most common gastrointestinal problem with felbamate,[31] vomiting, diarrhoea, nausea, abdominal pain and constipation may also occur, although possibly Drug Safety 1996 Dec: 15 (6)

387

Anticonvulsants in Childhood Epilepsy

no more often than with placebo.l 30] In one study, 7% of adults treated with tiagabine reported gastrointestinal complaints, significantly more than those given placebo.[ll] Childhood studies are not yet available for tiagabine. Pancreatitis has been reported in association with both carbamazepine[l2] and valproic acid therapy.lIO] With valproic acid, it can be acute and haemorrhagic and seems more common in children than adults,[IO] but is nevertheless very rare. A summary of gastrointestinal disturbances associated with anticonvulsant therapy is given in table VII. 4.6.3 Alterations in Appetite and Bodyweight

Excessive bodyweight gain may occur with valproic acid,[IO.IS] vigabatrin[22.4 9] or a combination ofthese 2 drugs.[SO] In one childhood study, 14% of those taking valproic acid were affected.[IS] Bodyweight gain can be sufficient reason to discontinue valproic acid therapy, particularly in adolescent girls.l IO ] Although excessive gains in bodyweight have been described in only 4.5% of one cohort of children given vigabatrin[49] - and this was considered to be less of a problem than with valproic acid[22]- 55% of children had excessive increases in bodyweight when vigabatrin was given to children already receiving valproic acid.1 s0 ] Anorexia can be a problem in 1% of children treated with carbamazepine,[17] and appeared to be responsible for bodyweight loss in 5% of those receiving valproic acid.[IS] Bromides may also be associated with reduced appetite.l 32 ] Bodyweight

loss has also been noted in 4% of patients receiving felbamate.l II ,30] In summary, adverse effects on appetite are most common when valproic acid and vigabatrin are given together, but of the 2 drugs these adverse effects are probably more of a problem with valproic acid than vigabatrin. 4.7 Renol Adverse Effects

Adverse renal effects associated with anticonvulsants are extremely rare, and are usually components of a multi systemic reaction. Acute interstitial nephritis may occur with carbamazepine.l 12 ] Nephrolithiasis has been reported in 1.5% of adults given topiramate.l 26] There have been no studies in children at this time. 4.8 Cosmetic Adverse Effects

Phenytoin is more likely to cause unwanted cosmetic effects than phenobarbital, primidone, carbamazepine or valproic acid.l48 ] Gingival hyperplasia, which is to some extent related to dosage and oral hygiene, is experienced by 25%,[15] and hirsutism in 9%, of children taking phenytoin.l 15 ] Hair loss precipitated by valproic acid[IO] has been recorded in 0.8% of children.[IS] 4.9 Adverse Effects on the Nervous System

The incidences of neurological disturbances, neurobehavioural changes and behavioural effects are summarised in table VIII. In the absence of closely controlled plasma concentrations, it is

Table VII, Incidence of gastrointestinal adverse effects (including pancreatitis) associated with anticonvulsant drugs Adverse effect

PBT

PRI

PHE

CBZ

OXCBZ

VAL

ETH

LTG

FBM

TGB

Mild disturbances

++

++

++

++

++

+++

+++

++

++

++

Constipation Eosinophilic colitis

+ VR

Vomiting Pancreatitis Excessive bodyweight gain

VGB

+

VR

VR

+++

++

Anorexia/bodyweightloss

+ + + Abbreviations and symbols: CBZ =carbamazepine; ETH =ethosuximide; FBM =felbamate; LTG =lamotrigine; OXCBZ =oxcarbazepine; PST = phenobarbital (phenobarbitone); PHE = phenytoin; PRI = primidone; TGB = tiagabine; VAL =valproic acid (sodium valproa\e); VGB =vigabatrin ; VR =very rare (single case reports); + =1.00 to 4.99%; ++ =5.00 to 9.99%; +++ =>10%.

© Adis International Limited. All rights reserved.

Drug Safety 1996 Dec; 15 (6)

Wallace

388

Table VIII. Incidence of neurological disturbances and neurobehavioural changes (at therapeutic dosages) associated with anticonvulsant drugs Effect

PBT

Neurological disturbances Headaches +

PAl

PHE

VAL

CBZ

OXCBZ ETH

VGB

LTG

BZD

GBP

FBM

TGB

TPA

BAa

+

+

+

++

+

+

±

+

++

+

+

+

+

+++

+ + +

+

++

Diplopia Tremor Ataxia at drug initiation

++

+

+++

Choreoathetosis

+ +

Dystonia

±

Asterixis Neurobehavioural changes ++a Somnolence/sedation Insomnia

++a

+

+

+ +

++

±

± ± ++a

+

++a

++

+

+

+

++

++

++

'Nervousness'

+++ +++

++

±

Irritability

+++

Oppositional behaviour

+++

+++

++

+

Aggression

+

+

++

Psychosis a

+

++

Weakness/tiredness Overactivity

++

+++ ++

± ± ++

+ + ++

+

±

Mainly at the start of treatment.

Abbreviations and symbols: BAa =bromides; BZD =benzodiazepines; CBZ =carbamazepine; ETH =ethosuximide; FBM =felbamate ; GBP =gabapentin; LTG =lamotrigine; OXCBZ =oxcarbazepine; PBT =phenobarbital (phenobarbitone) ; PHE =phenytoin; PAl =primidone; TGB =tiagabine; TPA =topiramate; VAL =valproic acid (sodium valproate); VGB =vigabatrin; ± =0.001 to 0.99%; + = 1.00 to 4.99%; ++ = 5.00 to 9.99%; +++ =>10%.

difficult to be sure that phenytoin causes more motor disturbances than other anticonvulsants. Certainly, other drugs can be implicated. 4.9. 1 Neurological Disturbances Headache

Headache is a common complaint in adults treated with anticonvulsants, and, as with other subjective phenomena, is probably under-reported in childhood. For example, in 1 study although 20% of all patients (children and adults) given felbamate complained of headaches,[30] in another study less than 5% of children seemed to be affected)31] Headache is reported to be one of the more common adverse effects of ethosuximide,[13] being recorded in 14% of children,[15] the same incidence as in children taking carbamazepine.[I5] Headache was noted in less than 1% of those children given lamotrigine)29] Comparable figures are not available for children, but 14% of adults given gabapentin reported headaches.[ 45 1 © Adis Internationa l Limite d . All rights reserved.

Diplopia

Diplopia is almost always associated with carbamazepine,[15.48,51] but may also occur with oxcarbazepine,[51] and, in adults, with gabapentin)45] Carbamazepine caused diplopia in 5%[15] and 3%[48] of children. Addition of lamotrigine to carbamazepine or oxcarbazepine therapy can precipitate diplopia via a pharmacodynamic interaction.[5l] Tremor

Tremor, similar to essential tremor, is particularly common with valproic acid:[IO,15] 10% of all patients,llOl and 15% of children,115] are reported to be affected. Sometimes valproic acid-induced tremor is dose-related) 10] Tremor that did not appear to be related to intention was noted in 14% of children treated with phenytoin,[15] I % of children taking carbamazepine,[15]9% of adults given tiagabine[25] and 12.5% of adults receiving gabapentin)45] Ataxia

Ataxia is well recognised as a complication of high plasma concentrations of many anticonvulsants, Drug Safe ty 1996 Dec; 15 (6)

Anticonvulsants in Childhood Epilepsy

but may occur transiently on initiation of treatment with phenobarbital, phenytoin, carbamazepine and, in particular, primidone.[l6] Temporary ataxia was reported in 4% of children given carbamazepine,l17] Although phenytoin was found to cause motor problems in childhood more often than phenobarbital, primidone, carbamazepine or valproic acid,[48] up to 50% alterations in phenytoin concentrations did not affect motor performance when this was measured objectively.[S2] On the other hand, impaired motor speed has been noted with higher plasma valproic acid concentrations.[S3] In addition, in a paediatric study, 6% of children given lamotrigine became ataxic)29] Other Motor Disabilities

Choreoathetosis has been reported in 2% of children taking phenytoin,[lS] and dystonia in 1% those receiving carbamazepine.112.1S] Asterixis has been linked with carbamazepine[12] or valproic acid therapy.[lO]

389

Dizziness and Asthenia

Dizziness and asthenia figure prominently in adult studies of new anticonvulsants,[2S,4S] but do not appear to be complaints that children report readily. Thus, their frequency is difficult to estimate. Early bromism is associated with complaints of weakness and tiredness.[32·s4] Nervousness

Nervousness is both difficult to specify and to quantitate in childhood, but has been noted in 12% of children treated with ethosuximide,[lS] and in 2% (compared with 19% with placebo) of patients given felbamate.[30] 4.9.3 Behavioural Effects

Overall, phenobarbital is considered to cause more behavioural problems than phenytoin, carbamazepine or valproic acid,[lS,48] more conduct disorders than carbamazepine[SS] and a higher incidence more overactivity, inability to stop, unhappiness and failure to finish tasks than valproic acid,ls6]

4.9.2 Neurobehavioural Changes

Overactivity Somnolence, Sedation and Coma

The earlier anticonvulsants all had sedative, as well as anticonvulsant, effects. Sedation is most likely to occur on initiation of treatment with primidone or carbamazepine,[16] but has also been reported in 13 % of children after introduction of ethosuximide therapy)13] Dosage-related sedation may be present in 2% of children treated with valproic acid,[10,48] and can occur with phenytoin, even at non-toxic plasma concentrations,l48] Sedation is a feature of bromism (excessive use of bromides).[S4] Drowsiness was a problem with carbamazepine in II % of children in one study.[48] Without specification of co-medication, vigabatrin caused sedation in 9% of children[49] and, when added to valproic acid, in 10%.[50] Although 6% of children given felbamate became sleepy, 5% developed insomnia. [31] Rare episodes of valproic acid-induced coma, which are not related to drug interactions or hyperammonaemia,[lO] are considered to be a result of underlying inborn errors of metabolism. © Adis International Limited. All rights reserved.

Overactivity attributable to anticonvulsants was first described with phenobarbital in 1955;[57] 39% of children in another study given phenobarbital had increased activity levels,llS] Those with preexisting overactivity are most likely to be affected.[S7] Children taking valproic acid are less likely to be overactive than those taking phenobarbitaJ.l48,S6] Overactivity may also occur in association with phenytoin,[48] benzodiazepines[lS] and gabapentin)S8.s9] Vigabatrin produced overactivity in 15%[50] and 26%[49] of childhood recipients, and this was particularly common in those with preexisting psychiatric problems,[49] or when dosages were higher. 122 ] Aggression and Irritability

Aggression, particularly in outbursts, has been reported in children treated with vigabatrin,[49] valproic acid,[lS] ethosuximide[lS] and gabapentin)S8,S9] Irritability has been described with phenobarbital,[60] valproic acid,ll5] benzodiazepines l15 ] and gabapentin.[59] Drug Safety 1996 Dec; 15 (6)

390

Sleep Disturbances

Increased wakefulness, can occur in children treated with phenobarbital,(60) phenytoin,(48) valproic acid,(15) ethosuximide(1 5) and vigabatrinJ22,49) Oppositional Behaviour

Oppositional behaviour has been reported in studies of gabapentin,158,59) but 'is also a feature of treatment with phenobarbitaJ.l55,56,60) Psychosis

Psychosis is a recognised complication of vigabatrin use in adults, and has now been reported in children,161] Psychotic episodes, which are extremely rare, have been noted in children, as well as in adults, given ethosuximide)13) Depression

Depression is unlikely to be a childhood complaint, but was found to be significantly more common (5%) in adults treated with tiagabine than in those receiving placeboJ25) 4.9,4 Cognitive Effects

Studies on the cognitive effects of anticonvulsants in children with epilepsy are complicated by frequent pre-existing learning difficulties, the effects of both clinical and subclinical seizures and problems with attention and concentration, in addition to the possible adverse effects of the anticonvulsant. In children with febrile seizures, higher plasma phenobarbital concentrations correlated with significantly reduced IQ scores, but the overall differences between phenobarbital and placebo were not significant)60] In similar children, significant changes in development were not found in those who received phenobarbital, when compared with valproic acid or no therapy.(62) However, in 28 older children, a double-blind crossover study showed that performance and full-scale intelligence scores were lower during phenobarbital, than during valproic acid, treatmentJ56) Nevertheless, in 73 children with newly diagnosed epilepsy, no significant difference was found on cognitive testing between those receiving phenobarbital, valproic acid or carbamazepineJ63) © Adis International Umited. All rights reserved.

Wallace

The American Academy of Paediatrics Committee[64] has reported that valproic acid had no effect on psychological performance or cognition; moreover, on detailed testing, improvements in focal attention, cancelling and the digit symbol substitution test have been reported with valproic acid. (65 ) A strictly monitored study of possible cognitive effects of phenytoin and carbamazepine found no significant changes on addition or withdrawal of the anticonvulsant,(66) although memory impairment,(67) reduced abilities on the peg board test and reductions in sustained attention (65 ) have been reported during treatment with carbamazepine, Complaints of poor memory and lack of concentration are features of early bromism. (32 ) It is clear that interference with cognitive processes is by no means confirmed for any drug that has been used in childhood epilepsy, when appropriate testing has been carried out. Comparative statements are, therefore, difficult. However, there is a suggestion that treatment with phenobarbital may be less desirable than with other anticonvulsants, Appropriate cognitive testing during treatment with recently developed anticonvulsants is awaited, The incidences of various cognitive effects associated with anticonvulsants are summarised in table IX,

5. Exacerbation of Seizures There is a tendency for reports on drug trials to emphasise the positive and underplay the negative findings. Thus, the potential for exacerbation of seizures may not be well publicised, It is often difficult to make out whether the underlying seizure condition has worsened or whether the anticonvulsant has exacerbated the problem. However, there seems no doubt that carbamazepine, vigabatrin, and probably lamotrigine and benzodiazepines, can increase the frequency of specific seizure types. When carbamazepine is introduced, atonic seizures may dramatically increase in number,[68] or appear for the first time.[69) Carbamazepine may also precipitate myoclonic seizures, and increase Drug Safety 1996 Dec; 15 (6)

391

Anticonvulsants in Childhood Epilepsy

Table IX. Incidence of various cognitive effects in clinical studies assessing the effects of anticonvulsant drugs Effect

PBT

Reduced IQ scores

±a

Improvements in specific tests

VAL

eBl

±

Reduced ability: peg board test

±

Reduced sustained attention

±

Poor memory/poor concentration a

Bromides

+

lamotrigine. Worsening of partial epilepsies was noted in 2 of 7 children given lamotrigine in an uncontrolled studyJ721 Of the benzodiazepines, clonazepam has occasionally been reported to increase myoclonia.[73] In summary, drugs acting via GAB A-related mechanisms and those causing sodium channel blockade may exacerbate seizures or precipitate new seizure types. In particular, myoclonia may appear or worsen.

Not significant when compared with placebo, VAL or eBl.

=

=

Abbreviations and symbols: eBl carbamazepine; PBT phenobarbital (phenobarbitone); VAL = valproic acid (sodium valproate); ± 0.001 to 0.99%; + 1.00 to 4.99%.

=

6. Conclusion

=

the frequency and duration of atypical absence seizures and generalised convulsive seizures.[68,69] Children who already have several seizure types are considered to be most at risk. When assessed on the basis of epileptic syndromes, children with nonprogressive myoclonic epilepsies were those most likely to have an increase in seizures after starting vigabatrin.[49] Increases were seen in a minority of children given vigabatrin for simple partial, tonic-clonic, atonic, atypical absence and myoclonic seizuresJ49] Furthermore, myoclonic seizures may present, de novo, during vigabatrin treatment.[70] Overall, an increase in seizures occurred in approximately 10% of 194 children given vigabatrinJ71] Seizures were most likely to increase in number in children with the Lennox-Gastaut syndrome or nonprogressive myoclonic epilepsy. Symptomatic infantile spasms, symptomatic partial and nonprogressive myoclonic seizures were most often reported to emerge as new seizure types after initiation of vigabatrin. [71] A greater than 50% increase in seizure frequency has b~en reported in 6% of children given lamotrigine.[291 The seizure types involved were nonprogressive myoclonic seizures, infantile spasms, other symptomatic generalised seizures, partial cryptogenic and partial symptomatic seizures. Two personal cases of severe myoclonic epilepsy in infancy were considerably worsened by © Adis International Limited. All rights reserved.

A very wide range of adverse effects may be observed in children with epilepsy who receive anticonvulsants. The nature of recording of responses to therapy in childhood, and the absence of placebocontrolled trials for most drugs, makes the incidence of adverse effects difficult to quantify. From the evidence available, life-threatening episodes appear most likely to occur with valproic acid, particularly when there is already an underlying disorder of mitochondrial function, and with felbamate, for reasons which are not yet understood. Minor, but uncomfortable, neurological and neurobehavioural disorders are the most prevalent adverse effects seen with anticonvulsants in children, with phenobarbital and vigabatrin causing most problems. To date however, there is no evidence that any anticonvulsant drug causes important cognitive problems. The rates of therapy discontinuation because of adverse events provide a useful comparison:[48] phenytoin 10%, valproic acid 8%, primidone 8%, phenobarbital 4% and carbamazepine 3%. In comparison, felbamate was discontinued in 10% of childrenpl vigabatrin in about 3 to 4%,[22,74,75 1and lamotrigine in just over 10%)76]

References I. Shaw PN, Houston JB, Rowland M, et al. Antipyrine metabolite kinetics in healthy human volunteers during multiple dosing of phenytoin and carbamazepine. Br J Clin Pharmacol 1985; 20: 611-8 2. Perucca E, Richens A. General principles: biotransformation In: Levy RH, Mattson RH , Meldrum BS. editors. Amiepileplic Drugs. 4th ed. New York: Raven Press, 1995: 31-50

Drug Safety 1996 Dec: 15 (6)

392

3. Connell J, Oozeer R, De Vries L, et al. Clinical and EEG responses to anticonvulsants in neonatal seizures. Arch Dis Child 1989; 64: 459-64 4. Bruni J. Phenytoin: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed. New York: Raven, 1995: 345-50 5. Cramer JA, Mattson RH. Phenobarbital: toxicity. In Levy RH, Mattson RH. Meldrum BS, editors. Antiepileptic Drugs. 4th ed. New York: Raven, 1995: 409-20 6. Schiff D, Chan G, Stern L. Fixed drug combinations and the displacement of bilirubin from albumin. Pediatrics 1971; 48: 139-41 7. Deshmukh A, Wittert W, Schnitzler E, et al. Lorazepam in the treatment of refractory neonatal seizures. Am J Dis Child 1986; 140: 1042-4 8. McDermott CA, Kowalczyk AL, Schnitzler ER, et al. Clinical and laboratory observations: pharmacokinetics of lorazepam in critical ill neonates with seizures. J Pediatr 1992; 120: 479-83 9. Sinal SH, Crowe JE. Cyanosis, coughing and hypotension following intravenous administration of paraldehyde. Pediatrics 1976; 57: 158-9 10. Dreifuss FE. Valproic acid: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed. New York : Raven, 1995: 641-8 II. Leppik I. Felbamate. Epilepsia 1995; 36 Suppl. 2: S66-72 12. Holmes G. Carbamazepine: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed. New York: Raven, 1995: 567-79 13. Dreifuss FE. Ethosuximide: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed. New York: Raven, 1995: 675-9 14. Leppik IE, Cloyd Jc. Primidone: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed. New York: Raven, 1995: 487-90 15. Abu-Arafeh lA, Wallace SJ. Unwanted effects of antiepileptic drugs . Dev Med Child Neurol 1988; 30: 117-21 16. Mattson RH. Efficacy and adverse effects of established and new antiepileptic drugs. Epilepsia 1995: 36 Suppl. 2: S13-26 17. Okuno T, Ho M, Nakano S, et al. Carbamazepine therapy and long-term prognosis in epilepsy of childhood. Epilepsia 1989; 30: 57-61 18. Lee WM . Drug-induced hepatotoxicity. N Engl J Med 1995; 333: 1118-27 19. Loiseau P. Sodium valproate, platelet dysfunction and bleeding. Epilepsia 1981; 22: 141-6 20. Gidal B, Spencer N, Maly M, et al. Valproate mediated disturbances of hemostasis [letter]. Neurology 1994; 30: 354 21. Allarakhia IN, Garofalo EA, Komarynski MA, et al. Valproic acid and thrombocytopenia in children: a case-controlled retrospective study. Pediatr Neuro11996; 14: 303-7 22. Appleton R. The role of vigabatrin in the management of infantile epileptic syndromes. Neurology 1993; 43 Suppl. 5: S21-3 23. Brodie MJ. Lamotrigine. Lancet 1992; 339: 1397-400 24. Ramsay RE. Gabapentin: toxicity. In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic Drugs. 4th ed. New York: Raven, 1995: 857-60 25. Leppik IE. Tiagabine: the safety landscape. Epilepsia 1995; 36 Suppl. 6: S 10-3 26. Reife RA, Pledger Gw. Safety of top iramate in clinical use. Adv AED Ther 1995; I (I): 24-9 27. Tada H, Wallace SJ, Hughes IA. Height in epilepsy. Arch Dis Child 1986; 61: 1224-6

© Adis Internatio nal Limited. All rights reserved.

Wallace

28. Dam M, Ostergaard LH. Other Antiepileptic drugs: oxcarbazepine. In: Levy RH, Mattson RH, Meldrum BS. editors. Antiepileptic Drugs. 4th ed. New York: Raven, 1995: 987-95 29. Schlumberger E, Chavez F, Palacios Let al. Lamotrigine in treatment of 120 children with epilepsy. Epilepsia 1994; 35: 359-67 30. Dodson WE. Felbamate in the treatment of Lennox-Gastaut Syndrome: results of a 12-month open-label study following a randomized clinical trial. Epilepsia 1993; 34 Suppl. 7: S18-24 3 I. Bebin EM, Dogis TF, Dreifuss FE. Felbamate: toxicity. In: Levy RH , Mattson RH, Meldrum BS, editors. Antiepileptic drugs. 4th ed . New York: Raven. 1995: 823-7 32. Dreifuss FE. Other antiepileptic drugs : bromides In: Levy RH. Mattson RH, Meldrum BS, editors. Antiepileptic drugs . 4th ed. New York: Raven, 1995: 949-51 33. Pellock JM. Carbamazepine side effects in children and adults. Epilepsia 1987; 28 Suppl. 3: S64-70 34. Jeng Y-M, Tien H-F, Su I-J. Phenytoin-induced pseudolymphoma: re-evaluation using modern molecular biology techniques. Epilepsia 1996; 37 (I): 104-7 35. Schaub JEM, Williamson PJ. Barnes EW. et al. Multisystem adverse reaction to lamotrigine [letter]. Lancet 1994; 344: 481 36. Baillie TA, Sheffels PR oValproic acid: chemistry and biotransformation . In: Levy RH, Mattson RH, Meldrum BS, editors. Antiepileptic Drugs. 4th ed. New York: Raven, 1995: 589-604 37. Melegh B, Pap M, Morava E, et al. Carnitine-dependent changes of metabolic fuel consumption during long-term treatment with valproic acid. J Pediatr 1994; 125: 317-21 38. Thurston JH, Hauhart RE. Reversal of the adverse chronic effects of the unsaturated derivative of valproic acid - 2-n-propyl-4-pentenoic acid - on ketogenesis and liver coenzyme A metabolism by a single injection of pantothenate, carnitine, and acetyl-cysteine in developing mice. Pediatr Res 1993; 33: 72-6 39. Murakami K, Sugimoto T, Nishira N, et al. Abnormal metabolism of carnitine and valproate in a case of acute encephalopathy during chronic valproate therapy. Brain Dev 1992; 14: 178-81 40. Kuhara T, Inoue Y, Matsumoto M, et al. Markedly increased ro-oxidation of valproate in fulminant hepatic failure. Epilepsia 1990; 31: 214-7 41. Konig St A, Siemes H, Blaker F, et al. Severe hepatotoxicity during valproate therapy: an update and report of eight new fatalities . Epilepsia 1994; 35: 1005-15 42. Bryant AE, Dreifuss FE. Valproic acid hepatic fatalities III: U.S. experience since 1986. Neurology 1996; 46: 465-9 43. Moore MR. International review of drugs in acute porphyria [letter]. Int J Biochem 1980; 12: 1089 44. Tatum WO, Zacharian SB . Gabapentin treatment of seizures in acute intermittent porphyria. Neurology 1995; 45: 1216-7 45. McLean MJ. Gabapentin. Epilepsia 1995; 36 Suppl. 2: S73-86 46. Dreifuss FE, Santilli N, Langer DH, et al. Valproic acid hepatic fatalities: a retrospective review. Neurology 1987; 37: 379-85 47. Scheffner D, Konig S, Rauterberg-Rutland I, et al. Fatal liver failure in 16 children with valproate therapy. Epilepsia 1988; 29: 530-42 48. Herranz JL, Armijo JA. Artego R. Clinical side effects of phenobarbitone primidone, phenytoin, carbamazepine and valproate during monotherapy in children. Epilepsia 1988; 29: 794-804 49. Dulac 0, Chiron C, Luna D, et al. Vigabatrin in childhood epilepsy. J Child Neuro11991; 6 Suppl. 2: S30-7

Drug Safety 1996 Dec: 15 (6)

Anticonvulsants in Childhood Epilepsy

50. Feucht M, Brantner-Inthaler S. y-vinyl GABA (vigabatrin) in the therapy of Lennox-Gastaut syndrome: an open study. Epilepsia 1994; 35 (5): 993-8 51. Alving 1. Case of severe acute intoxication with oxcarbazepine combined with lamotrigine [abstract]. Epilepsia 1994; 35 Suppl. 7: 72 52. Aman MG, Werry lS, Paxton lW, et al. Effects of phenytoin on cognitive motor performance in children as a function of drug concentration, seizure type, and time of medication. Epilepsia 1994; 35: 172-80 53. Aman MG, Werry lS, Paxton lW, et al. Effect of sodium valproate on psychomotor performance in children as a function of dose, fluctuations in concentration, and diagnosis. Epilepsia 1987; 28: 115-28 54. Steinhoff Bl, Kruse R. Bromide treatment of pharmaco-resistant epilepsy with generalised tonic-clonic seizures: a clinical study. Brain Dev 1992; 14: 144-9 55. Corbett lA, Trimble MR, Nichol RC. Behavioural and cognitive impairments in children with epilepsy: the long-term effects of anticonvulsant therapy. 1 Am Acad Child Psychiatry 1985; 24: 17-23 56. Vining EPG, Mellits ED, Dorsen MM , et al. Psychologic and behavioural effects of antiepileptic drugs in children: a double-blind comparative comparison between phenobarbital and valproic acid. Pediatrics 1987; 80: 165-74 57. Ounsted C. The hyperkinetic syndrome in epileptic children. Lancet 1955; II : 303-11 58. Wolf SM , Shinnar S, Hang H, et al. Gabapentin toxicity in children manifesting as behavioural changes. Epilepsia 1995; 36 (12): 1203-5 59. Lee DO, Steingard Rl , Cesena M, et al. Behavioural side effects of gabapentin in children . Epilepsia 1996; 37 (I): 87-90 60. Camfield CS, Chaplin S, Doyle AB, et al. Side effects of phenobarbital in toddlers; behavioural and cognitive aspects. 1 Pediatr 1979; 95: 361-5 61. Canovas Martinez A, Ordovas Baines lP, Escriva Asparisi A, et al. Vigabatrin - associated reversible acute psychosis in a child. Ann Pharmacother 1995; 29: 1115-7 62. Aldridge Smith J, Wallace Sl. Febrile convulsions: intellectual progress in relation to anticonvulsant therapy and recurrence of fits. Arch Dis Child 1982; 57: 104-7 63. Chen Y, Kang W, Chin-Min So W. Comparison of antiepileptic drugs on cognitive function in newly diagnosed epileptic children: a psychometric and neurophysiological study. Epilepsia 1996; 37 (I): 81-6

© Adis International Limited. All rights reserved.

393

64. American Academy of Pediatrics' Committee on Drugs. Behavioural and cognitive effects of anticonvulsant therapy. Pediatrics 1985; 76: 644-7 65. Stores G, Williams PL, Styles E, et al. Psychological effects of sodium valproate and carbamazepine in epilepsy. Arch Dis Child 1992; 67: 1330-7 66. Aldenkamp AP, Alpherts WC1 , Blennow G, et al. Withdrawal of antiepileptic medication in children - effects on cognitive function: the multicentre Holmfrid study. Neurology 1993; 43: 41-50 67. Forsythe I, Butler R, Berg I, et al. Cognitive impairment in new cases of epilepsy randomly assigned to carbamazepine, phenytoin and sodium valproate. Dev Med Child Neurol 1991; 33: 524-35 68. Shields WD, Saslow E. Myoclonic, atonic and absence seizures following institution of carbamazepine therapy in children. Neurology 1983; 33: 1487-9 69. Snead OC, Hosey LC. Exacerbation of seizures in children by carbamazepine. N Engl 1 Med 1985; 313: 916-21 70. Marciani MG, GigJi GL, Maschio CE, et al. Vigabatrin-induced myoclonus in four cases of partial epilepsy [abstract]. Epilepsia 1995; 36 Suppl. 3: 107 71. Lortie A, Chiron C, Mumford 1, et al. The potential for increasing seizure frequency, relapse and appearance of new seizure types with vigabatrin. Neurology 1993; 43 Suppl. 5: S24-7 72. Ceuleman B, Boel M. Lamotrigine: add on therapy in children with refractory epilepsies [abstract] . Epilepsia 1995; 36 Suppl. 3: 110 73. Roussounis SH, Rudolf N de M. Clonazepam in the treatment of children with intractable seizures. Dev Med Child Neurol 1977; 19: 326-34 74. Piatella L, Zamponi N, Cardinali C. Vigabatrin versus carbamazepine in newly diagnosed partial epilepsy in childhood [abstract] . Epilepsia 1995; 36 Suppl. 3: SI03 75. Herranz lL, Campos 1, Casas C, et al. Spanish experience with efficacy and tolerability of vigabatrin in 197 children with refractory epilepsy [abstract] . Epilepsia 1995; 36 Suppl. 3: SI03 76. Besag FMC, Wallace Sl, Dulac 0, et al. Lamotrigine in the treatment of epilepsy in childhood. Pediatr Pharmacol Ther 1995; 127: 991-7

Correspondence and reprints: Dr S.f. Wallace, Consultant Paediatric Neurologist, University Hospital of Wales, Heath Park, Cardiff CF4 4XW, Wales.

Drug Sofely 1996 Dec: 15(6)