Sorry, you need to enable JavaScript to visit this website.
Skip to main content

Antiseizure Medication Disorder

Brian K. Alldredge, PharmD | May 1, 2007
View more articles from the same authors.

Case Objectives

  • Appreciate the challenges of safe use of antiepileptic medications.
  • Identify phenytoin toxicity.
  • Understand the pharmacodynamics and pharmacokinetics of phenytoin.
  • List clinical scenarios when phenytoin levels may be unreliable.

Case & Commentary: Part 1

A 76-year-old man was admitted for evaluation of increasing lethargy, confusion, and decreased appetite. The patient had a past medical history of seizure disorder with a recent admission for uncontrolled seizures, anemia, and hip arthroplasty. His medications included phenytoin 300 mg once a day, phenobarbital 30 mg three times a day, risedronate 35 mg once a week, and iron supplements. On physical examination, his vital signs were unremarkable and he had no fever. By report, neurological exam revealed only confusion and diminished deep tendon reflexes. Laboratory data were significant for a slightly elevated white blood cell count (WBC) and abnormal urinalysis, with 6–10 WBCs. A phenobarbital level was 30 (therapeutic, 10–40 mcg/mL), and phenytoin level was 19 (therapeutic, 10–20 mcg/mL). The rest of his electrolytes, renal function, and liver function tests [aspartate aminotransferase (AST), alanine aminotransferase (ALT), bilirubin, and alkaline phosphatase] were normal. Computed tomography (CT) scan of the head was unremarkable. The providers thought that the patient's mental status change was likely due to urinary tract infection and effects of phenobarbital. The patient was treated with antibiotics, and his phenobarbital was held.

This patient—an elderly man taking phenytoin and phenobarbital chronically for a "seizure disorder"—probably has epilepsy, a disorder characterized by epileptic seizures that recur spontaneously. Epilepsy affects approximately 2 million people in the United States.(1) Antiseizure medications, which control seizures successfully in approximately two-thirds of patients, are the primary mode of treatment.(2) More than 15 antiseizure medications are commonly used in the United States. The ultimate drug choice is usually made based on patient-specific considerations, including predominant seizure type, drug–drug or drug–disease state interactions, and unique medication-related adverse effects.

Over the past 13 years, many new antiseizure medications have been introduced. Although most of them were initially approved for patients who fail to respond to "standard" agents, some of these newer drugs (e.g., oxcarbazepine, lamotrigine, topiramate) are being used as initial therapy in newly diagnosed epilepsy. In general, the newer medications appear to be equally effective in suppressing seizures as standard drugs and to have many fewer drug interactions than the older "enzyme-inducing" (e.g., phenytoin, carbamazepine, phenobarbital) or "enzyme-inhibiting" (e.g., valproate) drugs.(3) The clinical activity of the standard and new antiseizure drugs against various common seizure types is shown in the Figure. Ease of use is clearly a prominent factor for clinicians in the choice of antiseizure medication. For example, levetiracetam (Keppra), although not FDA-approved for use in newly diagnosed epilepsy patients, is emerging as a common first-choice agent in both hospital and ambulatory settings. Levetiracetam has the advantage of no significant drug-drug interactions, and, unlike most antiseizure drugs, which must be gradually titrated to an effective dose over several weeks, the starting dose of levetiracetam has been shown to effectively suppress seizures.

Antiseizure medications are not easy to use. In a recent study of hospitalized Medicare patients, antiseizure drugs were the seventh most common drug class associated with adverse drug reactions (ADRs)—an important finding given that only 3% of the 8 million inpatients studied were prescribed these medications for seizure disorders.(4) Antiseizure drugs are also used for select psychiatric disorders and for neuropathic pain management.

The safe prescribing and monitoring of antiseizure drugs are challenging for several reasons:

  • There is significant interpatient variability in the dosages needed to achieve the twin goals of freedom from seizures and tolerable adverse effects.
  • The therapeutic index (ratio between toxic dose and therapeutic dose, a measure of the relative safety of medications) of many antiseizure agents is low.
  • Many antiseizure medications have complex pharmacokinetics and drug interactions (particularly older drugs such as phenytoin, phenobarbital, and carbamazepine).

It is likely that all of these features contribute to the relatively high rate of ADRs with these drugs.

Most antiseizure medications are begun at low doses and gradually titrated upward based on clinical response. In patients with infrequent seizures, doses are increased to a serum concentration or daily dose that is generally effective. With older antiseizure drugs, the approach of targeting a serum concentration is usually favored, while a daily dose target is more useful with newer drugs because of the surprisingly poor correlation between serum concentrations of these agents and clinical response. For patients who need rapid protection from seizures, loading doses of select medications can be used, such as oral or intravenous phenytoin, intravenous valproate, or intravenous phenobarbital. Benzodiazepines can be given for near-immediate protection (e.g., intravenous lorazepam for status epilepticus) or as "bridge therapy" (e.g., scheduled oral doses of clonazepam for patients with seizures that have increased in frequency) until an effective oral maintenance dose can be attained. Examples of some of the more common dosing errors made with antiseizure medications are given in Table 1. Additional information on appropriate dosing of antiseizure drugs in adults is given in Table 2.

Case & Commentary: Part 2

Despite these interventions, the patient continued to be confused. A neurology consult was obtained. Review of the laboratory data revealed a serum albumin of 2.4 g/dL. This led to the calculation of a corrected phenytoin level: it was 33 (therapeutic, 10–20 mcg/mL). Phenytoin was held, and the patient's mental status returned to baseline in 72 hours.

Early clinical symptoms of dose-related phenytoin toxicity include dizziness, drowsiness, lethargy, and visual disturbances. As phenytoin levels continue to rise, ataxia and confusion may occur. Less commonly, patients experience an increase in the frequency or severity of seizures during phenytoin intoxication. In hindsight, this patient demonstrated some, but not all, of the classic symptoms of phenytoin toxicity. No mention is made of a medication-related change that might explain the elevated phenytoin level. When phenytoin toxicity is suspected, patients should be queried regarding recent dosage adjustments, changes in product appearance, and alterations in concomitant drug therapies that might influence phenytoin serum concentration (e.g., the addition of a phenytoin metabolism inhibitor such as cimetidine or the removal of a phenytoin metabolism inducer such as rifampin). As is clear from this case, a phenytoin blood level can be quite helpful even when there is no reason to suspect a change in phenytoin response.

Phenytoin is one of the drugs most commonly associated with preventable clinically significant adverse events.(5) Although the drug is highly effective, it has complex pharmacokinetics and requires careful monitoring and dosage adjustment, particularly in the elderly. The enzymes responsible for phenytoin metabolism (primarily CYP2C9 and CYP2C19) are saturable at drug concentrations near those used in clinical care. The consequence is that even small dosage changes can result in disproportionate changes in phenytoin blood levels and pharmacologic response. While equations can be used to calculate the dosing increment that would achieve a desired increase in phenytoin blood level, a simplified schema for adult patients is to limit dosage increases to 30–50 mg per day, and to increase dosages no more frequently than once every 1 or 2 weeks.

Some knowledge of phenytoin protein binding and the influence that protein binding variability has on phenytoin serum level interpretation is necessary to understand the error that occurred in this case. Phenytoin is approximately 90% bound to plasma proteins, primarily albumin. When considering a phenytoin drug level (such as "the therapeutic range of 10–20 mcg/mL"), it is important to realize that the level represents the total of both protein-bound drug and unbound phenytoin concentrations in serum. However, only unbound phenytoin has access to tissue sites that are responsible for phenytoin efficacy and toxicity.(6) Thus, a total (bound + unbound) phenytoin concentration value—the result reported by most clinical laboratories as the "phenytoin level"—is an indirect measure of the pharmacologically active (unbound) concentration.

Several factors can alter phenytoin protein binding and disrupt this normal 9:1 ratio of bound:unbound phenytoin (see Table 3). In the presence of these factors, phenytoin dosing need not be automatically altered; rather, the interpretation of the "serum phenytoin level" requires special consideration and care. Two common approaches are either to (i) order an unbound phenytoin level to see whether it falls within the unbound therapeutic range of 1–2 mcg/mL or (ii) use calculation methods to adjust the reported total phenytoin serum level to the value that would be expected if protein binding were not perturbed. I generally prefer to check the unbound phenytoin concentration, since calculation methods are not always accurate. For those who wish to use the second method, Table 4 shows how to correct a total phenytoin level for hypoalbuminemia. A different equation is used to correct phenytoin levels for end-stage renal disease, another common source of altered free phenytoin concentrations (Table 4).(7) Since an unbound phenytoin concentration was not reported in this case, the equation in Table 4 was likely used to estimate that this patient's phenytoin level (corrected for hypoalbuminemia) was approximately 33 mcg/mL. This finding led to the appropriate next step—holding phenytoin—which resulted in recovery of the patient's mental status.

Why did it take the patient 3 days to recover? This is not an unusual time course, since the rate of decline in phenytoin concentrations is often surprisingly slow when toxicity is due to high concentrations of unbound phenytoin. This phenomenon can be explained by saturable metabolism; at high serum levels, phenytoin half-life is much longer than it is at lower drug levels. Thus, the half-life of phenytoin not only varies between individuals, but it also varies within an individual depending on the serum concentration at any point in time. In this particular circumstance, the clinical dictum that "phenytoin half-life is about 24 hours" is inaccurate, which is why it took so long for the patient's mental status to recover. This is an important concept to recall—failure to do so might leave caregivers doubting their diagnosis of phenytoin toxicity on day 2 and embarking on inappropriate tests or treatments.

Given the complexity of this situation, one wonders whether a system-level solution might be helpful. Some institutions have in fact replaced total serum concentration monitoring with assays to measure only the unbound ("free") concentration of phenytoin, an intervention that would likely have prevented this case's error or more rapidly led to its recognition. Since such monitoring is more expensive and more time-consuming for lab personnel (an additional manual step is required to dialyze the sample across a filter membrane), one approach to targeting the use of unbound phenytoin monitoring would be to require that it be reported in patients with hypoalbuminemia or end-stage renal failure, which could be automatically triggered by a laboratory decision support engine. Since albumin concentrations are often low in hospitalized patients, another system-level intervention of potential value would be to require an albumin level determination within 30 days of any phenytoin level request. However, even in patients with no apparent alteration in protein binding (i.e., patients who lack any of the features listed in Table 3), unbound phenytoin concentrations more closely correlate with clinical status.(8) Thus, reporting unbound concentration may be preferred for all patients who require phenytoin monitoring.(9)

Another system intervention that can improve standardization of drug therapy—and has been shown to improve both economic and clinical outcomes with antiseizure drug therapy—is implementation of pharmacist monitoring and adjustment of these agents.(10) In 2003, 75% of U.S. states had enacted laws or practice act amendments to permit the enhanced involvement of pharmacists in direct drug therapy management, and, at present, the fastest growing clinical pharmacy service in U.S. hospitals is pharmacist-managed drug therapy.(10,11) Given a national shortage of pharmacists in the United States and their costs, even health care systems and hospitals that cannot afford widespread clinical pharmacist coverage might do well by considering targeted coverage in high-risk prescribing areas, such as antiseizure medications, pain medications, and anticoagulants.

Take-Home Points

  • Dosing and monitoring of antiseizure drugs are challenging due to the need to individualize therapy, the low therapeutic index of these agents, and complex pharmacokinetic characteristics of many first-generation drugs.
  • Newer antiseizure drugs are being used more commonly as initial therapy for epilepsy, partly because they are easier to use than older agents.
  • Phenytoin dosing and monitoring are complex due to saturable metabolism (leading to nonlinear pharmacokinetics) and the potential for displacement of active drug from albumin-binding sites.
  • In patients with low serum albumin or end-stage renal disease, total (bound + unbound) phenytoin levels can be misleading and must be interpreted with caution.
  • Unbound (free) phenytoin concentrations more closely correlate with clinical status of patients and are less likely to be misinterpreted than total phenytoin concentrations.
  • Pharmacist involvement in the management of antiseizure drug therapy can improve economic and patient outcomes.

Brian K. Alldredge, PharmD Professor of Clinical Pharmacy & Neurology Associate Dean, Academic Affairs, School of Pharmacy University of California, San Francisco

Faculty Disclosure: Dr. Alldredge has declared that neither he, nor any immediate member of his family, has a financial arrangement or other relationship with the manufacturers of any commercial products discussed in this continuing medical education activity. In addition, his commentary does not include information regarding investigational or off-label use of pharmaceutical products or medical devices.


1. Browne TR, Holmes GL. Epilepsy. New Engl J Med. 2001;344:1145-1151. [go to PubMed]

2. Duncan JS, Sander JW, Sisodiya SM, Walker MC. Adult epilepsy. Lancet. 2006;367:1087-1100. [go to PubMed]

3. Marson AG, Kadir ZA, Hutton JL, Chadwick DW. The new antiepileptic drugs: a systematic review of their efficacy and tolerability. Epilepsia. 1997;38:859-880. [go to PubMed]

4. Bond CA, Raehl CL. Adverse drug reactions in United States hospitals. Pharmacotherapy. 2006;26:601-608. [go to PubMed]

5. Winterstein AG, Hatton RC, Gonzalez-Rothi R, Johns TE, Segal R. Identifying clinically significant preventable adverse drug events through a hospital's database of adverse drug reaction reports. Am J Health Syst Pharm. 2002;59:1742-1749. [go to PubMed]

6. Soldin SJ. Free drug measurements. When and why? An overview. Arch Pathol Lab Med. 1999;123:822-823. [go to PubMed]

7. Winter ME. Phenytoin. In: Winter ME, ed. Basic Clinical Pharmacokinetics. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2004:321-363.

8. Kilpatrick CJ, Wanwimolruk S, Wing LM. Plasma concentrations of unbound phenytoin in the management of epilepsy. Br J Clin Pharmacol. 1984;17:539-546. [go to PubMed]

9. Burt M, Anderson DC, Kloss J, Apple FS. Evidence-based implementation of free phenytoin therapeutic drug monitoring. Clin Chem. 2000;46:1132-1135. [go to PubMed]

10. Bond CA, Raehl CL. Clinical and economic outcomes of pharmacist-managed antiepileptic drug therapy. Pharmacotherapy. 2006;26:1369-1378. [go to PubMed]

11. Hammond RW, Schwartz AH, Campbell MJ, et al. Collaborative drug therapy management by pharmacists—2003. Pharmacotherapy. 2003;23:1210-1225. [go to PubMed]


Figure. Clinical Activity of Antiseizure Drugs for Various Types of Epileptic Seizures. "Standard" antiseizure drugs shown in black; newer antiseizure drugs are shown in red.


Table 1. How to Avoid Common Errors in Dosing and Monitoring of Antiseizure Drugs

Error Comments and Avoidance Strategies
Initiating therapy at a dose that is too high or escalating the dosage too quickly Dose-related adverse effects often cause patients to reject a potentially useful medication. For example, an adult patient started on carbamazepine, 400 mg BID, might develop dizziness and refuse to continue therapy. Problems can usually be avoided if starting doses are not excessive (a more appropriate starting dose for carbamazepine would have been 200 mg BID) and if dosing escalations are made at 1- or 2-week intervals for outpatients (see Table 2). Intervals between dose escalations can be shorter for hospitalized patients since adverse effects can be more closely monitored and managed.
Not considering drug interactions when devising an initial dosing scheme Lamotrigine is a particularly important drug to consider in this regard. Starting doses vary significantly depending on whether the patient is taking other antiseizure medications. For example, a lamotrigine starting dose of 25 mg BID would be appropriate for a patient taking phenytoin concurrently. However, this same starting dose would be excessive in a patient taking valproate concurrently. Patients taking valproate should begin lamotrigine at a dose of 25 mg every other day to reduce the risk of a life-threatening rash.
Adjusting medication doses based solely on blood level results Therapeutic blood level ranges are useful as guides, but individual patients may attain therapeutic success (i.e., seizure control and good medication tolerability) at blood levels outside of these usual ranges. Problems can be avoided by adjusting doses based on seizure recurrence and symptoms of dose-related drug toxicity (e.g., dizziness, drowsiness, diplopia, and ataxia). The common clinical dictum is "treat the patient, not the number."
Withholding further dose escalations in patients with a partial response and good medication tolerability due to fear of exceeding the "usual" or "maximal" dose In adults, maintenance doses of antiseizure medications can vary fivefold or more, and doses need to be individualized based on clinical response. There are no a priori dosing restrictions for any of the commonly used antiseizure drugs. Drug doses can continue to be increased as long as the patient derives incremental benefit in seizure frequency (or severity) and is able to tolerate any medication-related adverse effects.
Escalating antiseizure medication doses too slowly in patients with uncontrolled seizures Rather than increasing medication doses in small increments at each follow-up visit, any antiseizure drugs can be adjusted by giving the patient a schedule to increase the dose at 1- or 2-week intervals until their "maximal tolerated dose" is reached (usually indicated by mild drowsiness or transient diplopia) and then instructing the patient to drop back to a dose that was well tolerated. This approach is reasonable in patients with frequent seizures. In patients with infrequent seizures, doses are usually titrated to a "usually effective dose" or a "usually effective serum concentration."

Table 2. Dosing of Select Antiseizure Drugs in Adults*

Drug Starting Dose (mg/day) Dosing Frequency Incremental Increases Usual Maintenance Dose (mg/day)
Carbamazepine 400 Standard release (e.g., Tegretol and generic): BID–TID Extended-release (e.g., Carbatrol and Tegretol-XR): BID 200 mg/day at weekly intervals 600–1800
Ethosuximide 500 BID 250 mg/day at 1-week intervals 1000–2000
Gabapentin 900 TID (QID for doses >3600 mg/day) 300–600 mg/day at 2- to 7-day intervals 900–3600
Lamotrigine (in patient on valproate alone) 12.5 (often given as 25 mg every other day) BID 25–50 mg/day at 1- to 2-week intervals 100–200
Lamotrigine (in patient on phenytoin, carbamazepine, phenobarbital, or primidone)** 50 BID 50–100 mg/day at 1- to 2-week intervals 300–500
Lamotrigine (in patient taking no other antiseizure drugs) 25 BID 50 mg at 1- to 2-week intervals 225–500
Levetiracetam 500 BID 500–1000 mg at 1-week intervals 1000–3000
Oxcarbazepine 600 BID 600 mg/day at 1-week intervals 600–2400
Phenytoin 300 Daily 30–50 mg/day at 1- to 2-week intervals 300–500
Topiramate 25–50 BID 25–50 mg/day at 1-week intervals 200–400
Zonisamide 100 QD–BID 100 mg/day at 2-week intervals 400–600
Valproate 500 Delayed-release (e.g., Depakote): BID Extended-release (e.g., Depakote-ER): QD 250 mg/day at 1-week intervals 1000–3000

*Patients 16 years and older.

**In patients taking valproate with either phenytoin, carbamazepine, phenobarbital, or primidone, the dosing scheme for lamotrigine in patients taking no other antiseizure drugs should be followed.

Table 3. Factors That Can Alter Phenytoin Protein Binding (4)
End-stage renal disease (creatinine clearance
Presence of displacing drugs (e.g., aspirin, valproic acid)

Table 4. Correcting a Total (Bound + Unbound) Phenytoin Level for Hypoalbuminemia (4)

Use the following equation, in which C’ is the observed total plasma concentration reported by the laboratory and P’ is the patient's albumin concentration (in g/dL). The resulting value (C normal binding) is the value that would be expected if the albumin concentration were normal.
Correcting a Total (Bound + Unbound) Phenytoin Level for Hypoalbuminemia and End-Stage Renal Disease (4)
This project was funded under contract number 75Q80119C00004 from the Agency for Healthcare Research and Quality (AHRQ), U.S. Department of Health and Human Services. The authors are solely responsible for this report’s contents, findings, and conclusions, which do not necessarily represent the views of AHRQ. Readers should not interpret any statement in this report as an official position of AHRQ or of the U.S. Department of Health and Human Services. None of the authors has any affiliation or financial involvement that conflicts with the material presented in this report. View AHRQ Disclaimers
Related Resources From the Same Author(s)
Related Resources