Tough seizures: When the first anticonvulsant fails


Children with epilepsy can usually achieve seizure control, but not always with the first medication prescribed. General pediatricians can learn from a child neurologist&s approach to this problem.


Tough seizures:
When the first anticonvulsant fails

Jump to:
Choose article section... What will the neurologist do first?

By Eric H. Kossoff, MD

Children with epilepsy can usually achieve seizure control, but not always with the first medication prescribed. General pediatricians can learn from a child neurologist's approach to this problem.


Your patient, a 6-year-old girl with infrequent generalized tonic-clonic convulsions but "staring and nonresponsive spells" multiple times each day, continues to have events despite high doses of carbamazepine. When you increase the dose even further she becomes ataxic and sedated. What should you do next?

Another patient, a 15-year-old girl with seizures for the past year, continues to have events despite high doses of valproic acid. All her events happen in the early morning on the school bus and have never been witnessed by her parents. Her electroencephalogram (EEG) has been normal three times and a magnetic resonance imaging (MRI) scan done the week before was also normal. What is your next step?

Last, a 9-year-old boy with daily seizures has not responded to high doses of carbamazepine that you have prescribed. His parents describe focal jerking of his left side with all seizures, and an EEG shows right temporal sharp waves consistently. An MRI scan is normal. How should he be treated now?


"We believe your child has epilepsy." Very little in pediatrics can be as terrifying to parents as that statement, often said after a child has had two or more seizures.1 The term "epilepsy" invokes thoughts of loss of control over all body functions and unpredictability. For thousands of years, patients with seizures were seen as demonically possessed. Over the past 150 years, however, the emergence of both the EEG and anticonvulsants has revolutionized our understanding of epilepsy and our ability to treat it. The growth in the number of available anticonvulsants, including a dramatic increase in options during the past decade (Figure 1), has enabled most children with epilepsy to obtain seizure control without any interference in normal home and school functions.



Despite so many options, the decision to start anticonvulsant therapy remains a complicated one. After a single, unprovoked, nonfebrile seizure, the risk of recurrence is approximately 30%.2 Once a second seizure occurs, the risk of a third seizure more than doubles—to 70%.2,3 However, even with this increase in incidence and even after two seizures, many variables—including patient compliance, severity and frequency of seizures, epilepsy syndrome if present, and, of course, parental preference—influence the decision to treat.

The choice of many parents to defer use of an anticonvulsant is an understandable one. All anticonvulsants can cause side effects, both modest (sedation, rash, ataxia) and severe (Stevens-Johnson syndrome, hepatotoxicity). The ability to drive a motor vehicle safely is a consideration in the decision to use an anticonvulsant in an adolescent, but it is not an issue with a child. In addition, most children are well supervised, especially when bathing or swimming— although pediatricians should still remind parents of the need for close supervision because of the possibility of another seizure. Moreover, seizures, while bothersome and occasionally prolonged, do not cause brain damage or learning disability.4,5

When anticonvulsant therapy is started, most children respond to the first agent chosen. In adults, approximately 50% respond to the first drug with total seizure control, 90% of whom do so with only a modest dosage.6 Child neurologists often pick anticonvulsants based on the child's age, seizure type (generalized vs. partial), side effect profile, and epilepsy syndrome if known (for example, absence, benign rolandic, juvenile myoclonic epilepsy). The EEG can be very influential in this regard.7,8 Frequently, children can be tapered off medication after several years without seizures, as many will "outgrow" their seizures. That is especially true with certain epilepsy syndromes, such as benign rolandic and absence epilepsy.

A small percentage of children with epilepsy are described as "intractable." The definition of this varies: from failure to respond to one drug to failure to respond to multiple agents.9 Pediatricians are certainly capable of managing children who have had only a few seizures, especially if they have responded successfully to the first anticonvulsant prescribed. If, however, a child begins to have breakthrough seizures—even after doses of the antiepileptic drug have been maximized—the primary care pediatrician should refer the patient to a child neurologist or epileptologist. The experience and training of these subspecialists can be invaluable in the decision-making process.

What will the neurologist do first?

A neurologist who is asked to evaluate a child with epilepsy after the first medication has failed should, before anything else, confirm three things: that the child in fact has epilepsy; that the anticonvulsant chosen makes sense based on the individual child and his or her seizures; and, last, that the anticonvulsant actually failed. An algorithm for approaching these patients is presented in Figure 2.



Is it in fact epilepsy? It is always prudent for the specialist to confirm that the child actually has epilepsy. It is not uncommon for a child who has been on medications for several years to be referred to a child epileptologist from a primary care provider or even a child neurologist, only for the epileptologist to discover that the reason the drugs have not worked is because the child does not have epilepsy! The differential diagnosis for nonepileptic seizures is large and varies somewhat depending on the child's age (Table 1). Repeatedly normal EEGs, events not witnessed by health professionals or parents, and consistent triggers for events (such as sleep or stress) should raise a red flag for a nonepileptic cause. In many cases, anticonvulsants can actually worsen the patient's condition—as with sleep myoclonus, breath-holding, staring spells, and syncope, for example—primarily because of their sedative side effects.


Major nonepileptic causes of seizure-like events


There is no substitute for a detailed history in potential nonepilepsy situations. Nonetheless, combination video-EEG monitoring over a prolonged period, though expensive, is often necessary to make the diagnosis. These children may require several days of monitoring in the hospital after medications have been withdrawn to "capture" events. At the least, families can try to video record an event; many families have access to a portable camcorder. A picture is worth a thousand words, as the saying goes, but a videotape can be worth a million. Have the parents send a videotape to your office.

Is it the right anticonvulsant? Assuming the child is indeed having seizures, the choice of anticonvulsant should be carefully evaluated. Does the child have an epilepsy syndrome that was missed? Benign rolandic epilepsy (nocturnal seizures, often starting in the face unilaterally, with central-temporal spike waves on EEG) is often exquisitely sensitive to carbamazepine and less so to other anticonvulsants. Absence epilepsy (formerly known as petit mal) usually responds to ethosuximide or valproic acid but may be worsened by carbamazepine.10 Table 2 lists special indications for the various anticonvulsants.



Is the seizure more likely to be generalized (entire brain involved at onset) or partial (one part of the brain)? An initially normal EEG may lead a neurologist to choose a partial epilepsy agent, because most cases of generalized epilepsy have an abnormal interictal (between seizures) EEG. Not all anticonvulsants are equally effective in treating the different epilepsy types, however, and an anticonvulsant can occasionally make another type of seizure pattern worse.

Did the drug actually fail? One of the most common histories provided to child epileptologists is that the child failed multiple anticonvulsants. However, a detailed drug history reveals that some of these agents were tried for only several days and discontinued because of perceived side effects or lack of efficacy. Most anticonvulsants need several weeks to show their effectiveness.

Another common mistake is considering a drug a failure if the serum level is in the therapeutic range but there is no seizure control. Because therapeutic levels were established based on limited data for 95% of adults in initial clinical trials who had obtained seizure control, the usefulness of a level is limited. Levels can, however, be useful for measuring compliance or as a general marker for determining the serum level at which control was obtained in that individual child. Anticonvulsants should be titrated up to the level required for seizure control, not just to a "therapeutic" level; the occurrence of side effects should serve as the limiting factor. No agent should be discarded as failed until the dose has been maximized. Each child has a unique metabolism and may require double or even triple the recommended dose per kilogram of body weight to achieve seizure control.

As with any medical condition, adherence can be an issue. Children or adolescents responsible for taking their own medications should be specifically asked, occasionally with their parents out of the examining room, if they are actually taking their medicine. Obviously, no anticonvulsant will work if it is not taken.

In cases where patients are nonadherent, are side effects the major reason? Sometimes an anticonvulsant can make a child seizure free but the cost to the child's well- being can make it intolerable. If so, a change may be in order—as in the case of an overweight adolescent female who experienced a 20 lb weight gain and hair thinning after being placed on valproic acid. All anticonvulsants, especially the older agents, have predictable, often dose-dependent, side effects, which are listed in Table 2.

Fluctuations in the anticonvulsant serum level are a potentially amenable variable that can be adjusted to make an agent more efficacious and enhance compliance. Twice daily dosing is much more convenient for families than dosing three or four times a day. Although the half-life of many agents is six to eight hours, seizures in children with uncomplicated epilepsy can often be controlled with bid dosing—tid or qid dosing is rarely necessary. A child with a history of seizure exacerbations before a dose is due to be given, however, should be switched to more frequent dosing or to an extended-release formulation (available for carbamazepine, phenytoin, and valproic acid) before the anticonvulsant is considered to have failed.

Add on? Or substitute?

Once an anticonvulsant has been tried and deemed ineffective, the decision to either add a second agent or to make a substitution is not clear-cut. If the dosage of the first agent was escalated but seizure control was not achieved or intolerable side effects developed, that drug should without question be discontinued and a new drug substituted. Polypharmacy has its share of problems, including the increased risk of drug interactions (for example, the combination of valproic acid and lamotrigine causes rash; carbamazepine decreases levels of many other anticonvulsants), side effects, and teratogenicity.11 It is certainly tempting simply to continue to add drugs when a child is having persistent seizures, but it can be counterproductive in the long run. In a recent consensus guideline of epilepsy experts, approximately 90% thought that a second attempt at monotherapy (substitution) was preferable to add-on therapy.12 Even after a second drug failed, physician preference for trying a third agent or for using a combination of therapies was about equal.12

When a child has some positive response to the first medication but is experiencing the "law of diminishing returns," it might be beneficial to continue the drug at a dose that does not cause side effects and add a second agent for combined effect. A recent study in adults showed a slightly higher chance of improvement in seizure control, without a worsening of side effects, by adding a second agent (26% seizure free) compared with making a substitution (17%).13 When adding a second drug, it may be preferable to use an agent from a class whose mechanism of action is different from that of the original medication (Table 3).13 Specifically, using the combination of a sodium channel blocker and a drug with more than one mechanism of action may be ideal—such as lamotrigine and valproic acid, or carbamazepine and topiramate.13 In addition, certain drug combinations, such as lamotrigine and valproic acid, have a synergistic effect, making their sum greater than their parts.



The decision to slowly increase the dose of a new drug to a standard level before tapering the initial drug, rather than tapering it immediately, is a matter of style not science. Even if the first agent did not seem to have benefit, suddenly tapering it without having another anticonvulsant on board may precipitate an increase in seizures. Slowly tapering medications over four to six weeks seems to be the ideal.

An abundance of choices

Once the decision to try a new agent is made, the number of choices available can seem overwhelming. Over the past decade, the number of available anticonvulsants has doubled. This growth has both improved the options for children with epilepsy and made deciding which anticonvulsant to use more difficult.14

Older anticonvulsants, such as phenobarbital, carbamazepine, valproic acid, and phenytoin, have been used for decades, are well-studied, and can often be considerably less expensive than newer agents. They come in multiple formulations, including extended-release preparations, capsule sprinkles, chewable tablets, oral suspensions, and even intravenous solutions. For these reasons, most neurologists use these older drugs as first-line medications. On the other hand, many of the newer agents, such as gabapentin, oxcarbazepine, and lamotrigine, have fewer side effects and fewer drug interactions, and may have additional benefits for pain and migraine.15,16

Is surgery an option?

In one large study, approximately 50% of patients became seizure free with the first anticonvulsant chosen. But, when the drug failed, only 11% became seizure free (without intolerable side effects) with a second agent.17 With such discouraging results, the physician always needs to consider in the back of his or her mind whether a child who fails the first agent is potentially a candidate for surgery. Epilepsy surgery can range from relatively small resections such as lesionectomy to larger operations such as temporal lobectomies and hemispherectomies. Recent evidence in adults suggests that, after one year, surgery is superior to medical management when an epileptic focus is located in the temporal lobes.18 The study found that 58% of these surgical patients were seizure free at one year, compared with 8% in the medical group. Evidence from children and even infants is equally strong, with approximately 60% seizure free with surgery.19

Surgery should be strongly considered when several drugs have failed, there is a single epileptogenic focus, and that focus is located in a functionally silent cortex. Most epilepsy experts consider surgery after three antiepileptic drugs tried as monotherapy or at least one combination trial has failed.12 Resection of an epileptogenic focus can save a child from years of medications with potential side effects and can be curative rather than solely suppress seizures. New neuroimaging technologies and surgical techniques have improved outcomes in even the most difficult cases.

Nonpharmaceutical treatments

The following options should be reserved for children with intractable epilepsy who do not have a clear structural lesion amenable to surgical resection.

The ketogenic diet is a high fat, adequate protein, low carbohydrate diet that has been used for the treatment of intractable childhood epilepsy since the 1920s.20 The diet mimics the biochemical changes associated with starvation that have been well-documented to reduce seizures.21,22 Typical meals include eggs, bacon, heavy cream, tuna, and mayonnaise, without breads and sugars and only limited fruits and vegetables. Calories and fluids are carefully measured by weighing foods. Urine ketones are checked frequently to monitor efficacy. The ketogenic diet was used less often from 1940 to 1990 than previously because of the increased availability of anticonvulsants such as phenytoin and carbamazepine, but it has reemerged recently as a viable therapeutic option.23 Although difficult to adhere to, the diet can result in a more than 50% reduction in seizures in half of children with epilepsy that is hard to control.24 Well-established side effects include constipation, limited weight gain, and kidney stones (5%).

Vagal nerve stimulation (VNS). With VNS, approved for use in 1997, a small metal device is surgically implanted in the chest and wires are connected to the vagus nerve, which is then electrically stimulated periodically. It is postulated that this stimulation then travels toward the thalamus, causing a desynchronization of seizure discharges and thereby preventing seizures. VNS also has the benefit of allowing patients to abort seizures by holding a magnet over the device to induce stimulation. Although expensive, VNS has been shown to be efficacious for refractory epilepsy in children.25 Side effects include periodic voice hoarseness.

Parental involvement

As with any complex disorder in a pediatric patient, having a well-educated parent is a definite advantage. There are many books and World Wide Web resources available for parents and children to read, with easy-to-understand information (see "For patients and their parents: Resources about epilepsy"). In addition, the Epilepsy Foundation, through its local chapters, will advocate for children with mild to severe epilepsy who are having difficulty with school or any other issue.

With many new anticonvulsants (and nonpharmacologic therapies) available, a majority of children with epilepsy can have seizure control without compromising quality of life. Key points to remember in caring for children with this condition are listed in the box on the left.



The author gratefully acknowledges the assistance of Dr. Eileen P. Vining, Dr. John M. Freeman, and Heather D. Hladky, RN, in editing this manuscript before submission.



1. Commission on epidemiology and prognosis, International League Against Epilepsy: Guidelines for epidemiologic studies on epilepsy. Epilepsia 1993;34:592

2. Shinnar S, Berg AT, O'Dell C, et al: Predictors of multiple seizures in a cohort of children prospectively followed from the time of their first unprovoked seizure. Ann Neurol 2000;48:140

3. Hauser WA, Rich SS, Annegers JF, et al: Seizure recurrence after a first unprovoked seizure: An extended follow-up. Neurology 1990;40:1163

4. Musico M, Beghi E, Solari A, et al, for the First Seizure Trial Group (FIRST Group): Treatment of first tonic-clonic seizure does not improve the prognosis of epilepsy. Neurology 1997;49:991

5. Shinnar S, Berg AT: Does antiepileptic drug therapy prevent the development of "chronic" epilepsy? Epilepsia 1996;37:701

6. Kwan P, Brodie MJ: Effectiveness of first antiepileptic drug. Epilepsia 2001;42:1255

7. Shinnar S, Berg AT, Moshe SL, et al: The risk of seizure recurrence following a first unprovoked afebrile seizure in childhood: An extended follow-up. Pediatrics 1996;98:216

8. Hirtz D, Ashwal S, Berg A, et al: Practice parameter: Evaluating a first nonfebrile seizure in children: Report of the quality standards subcommittee of the American Academy of Neurology, The Child Neurology Society, and The American Epilepsy Society. Neurology 2000;55:616

9. Crumrine PK: Antiepileptic drug selection in pediatric epilepsy. J Child Neurol 2002;17:2S2

10. Guerrini R, Belmonte A, Genton P: Antiepileptic drug-induced worsening of seizures in children. Epilepsia 1998;39(suppl 3): S2

11. Ferrendeli JA: Pharmacology of antiepileptic drug polypharmacy. Epilepsia 1999;40(suppl 5):S81

12. Karceski S, Morrell M, Carpenter D: The expert consensus guideline series: Treatment of epilepsy. Epilepsy and Behavior 2001;2:A1

13. Kwan P, Brodie MJ: Epilepsy after the first drug fails: Substitution or add-on? Seizure 2000;9:464

14. Pellock JM, Appleton R: Use of new antiepileptic drugs in the treatment of childhood epilepsy. Epilepsia 1999;40(suppl 6):S29

15. Cutrer FM, Silberstein SD, Mathew NT, et al: Antiepileptic drugs in migraine, cluster headache, and mood disorders. Headache 2001;41(suppl 1):1

16. Mellegers MA, Furlan AD, Mailis A: Gabapentin for neuropathic pain: Systematic review of controlled and uncontrolled literature. Clin J Pain 2001;17:284

17. Kwan P, Brodie MJ: Early identification of refractory epilepsy. N Engl J Med 2000;342:314

18. Wiebe S, Blume WT, Girvin JP, et al: A randomized, controlled trial of surgery for temporal-lobe epilepsy. N Engl J Med 2001;345:311

19. Saneto RP, Wyllie E: Epilepsy surgery in infants. Seminar Pediatr Neurology 2000;7:187

20. Swink TD, Vining EPG, Freeman JM: The ketogenic diet: 1997. Adv Pediatr 1997;44:297

21. Lennox WG: Ketogenic diet in treatment of epilepsy. N Engl J Med 1928;199:74

22. Geyelin HR: Fasting as a method for treating epilepsy. Med Record 1921;99:1037

23. Lefevre F, Aronson N: Ketogenic diet for the treatment of refractory epilepsy in children: A systematic review of efficacy. Pediatrics 2000;105:e46

24. Freeman JM, Vining EPG, Pillas DJ, et al: The efficacy of the ketogenic diet-1998: A prospective evaluation of intervention in 150 children. Pediatrics 1998;102;1358

25. Valencia I, Holder DL, Helmers SL, et al: Vagal nerve stimulation in pediatric epilepsy: A review. Pediatr Neurol 2001;25:368

DR. KOSSOFF is assistant professor, neurology and pediatrics, The Johns Hopkins Hospital, Baltimore, Md.

Revisiting our three patients

Case 1. Based on the description of her events, this 6-year-old girl may have absence seizures. An EEG reveals three-per-second spike-and-wave discharges diffusely. You discontinue carbamazepine and begin valproic acid. Within two weeks, her staring spells vanish.

Case 2. The 15-year-old girl with events on the bus is brought in for video EEG monitoring. During her admission she has two events of full body tremors with apparent nonresponsiveness despite a normal waking EEG. These events are thought to be nonepileptic (pseudoseizures). Her valproic acid is discontinued and she is referred for psychological evaluation and therapy.

Case 3. The 9-year-old boy with stereotyped left focal seizures and right temporal abnormalities is admitted for video-EEG monitoring. His carbamazepine is tapered and he has four seizures, all of which originate from the right midtemporal lobe. His carbamazepine is restarted and lamotrigine is initiated. Further investigation (including positron emission tomography scan, magnetic resonance spectroscopy, and subdural grid electrode implantation) is considered, given that he is a potential surgical candidate if lamotrigine and, perhaps, a third agent fail.

For patients and their parents: Resources about epilepsy

Seizures and Epilepsy in Childhood: A Guide for Parents, by JM Freeman, EP Vining, and D Pillas, Johns Hopkins University Press, 1997

Epilepsy: Patient and Family Guide, by O Devinsky, FA Davis Co, 2001

The Brainstorms series of books, edited by SC Schachter, Lippincott-Raven

American College of Physicians Home Medical Guide: Epilepsy, edited by DR Goldmann and DA Horowitz, DK Publishing, 2000

Epilepsy Foundation
Physician and parent information

Epilepsy Circle of Support
(information and support for parents and children) Bergeron/epilepsyhome.html

Epilepsy information from the pharmaceutical company Novartis


When an anticonvulsant fails


Eric Kossoff. Tough seizures: When the first anticonvulsant fails. Contemporary Pediatrics 2002;9:85.

Recent Videos
Wendy Ripple, MD
Wendy Ripple, MD
Lawrence Eichenfield, MD
Lawrence Eichenfield, MD | Image credit: KOL provided
FDA approves B-VEC to treat dystrophic epidermolysis bullosa patients 6 months and older | Image Credit: bankrx - Image Credit: bankrx -
Related Content
© 2024 MJH Life Sciences

All rights reserved.