Antiepileptic Drugs, 5th Edition

General Principles


Laboratory Monitoring of Antiepileptic Drugs

Svein I. Johannessen PhD*

Torbjörn Tomson MD, PhD**

* Director of Research, The National Center for Epilepsy, Sandvika, Norway

** Department of Neurology, Karolinska Hospital, Stockholm, Sweden

Because drug action depends on drug disposition, knowledge of the fundamental pharmacokinetic properties of a given drug is important for optimal treatment. Drug treatment of epilepsy was one of the first areas to benefit from clinical pharmacokinetic studies. In the past, the most effective individual dosage was established by trial and error. The development of technology for quantifying drug concentrations in biologic fluids has, however, made it possible to study the relationship among drug dosage, drug concentration in body fluids, and pharmacologic effect and thereby provides new insight into drug therapy (1,2). It was soon recognized that the desired therapeutic effect of many antiepileptic drugs (AEDs) was achieved within specific ranges of serum levels for each drug: lower concentrations gave an unsatisfactory effect, and higher concentrations gave undesirable side effects. Therapeutic drug level monitoring has since become established for many drugs to optimize drug therapy regimens for individual patients.

Mainly because of indiscriminate overuse and misuse, therapeutic drug monitoring has attracted criticism (3,4), and its value in the treatment of epilepsy has sometimes been questioned. However, although randomized clinical trials have failed to demonstrate an impact of drug monitoring on the overall outcome of epilepsy treatment (5), better understanding of the relationship between pharmacokinetics and pharmacodynamics is likely to result in improved drug effectiveness and safety (6). Numerous reports have been published on the kinetics of AEDs and on the rational use of drug monitoring (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16).


The role of therapeutic drug monitoring in the treatment of epilepsy reflects epilepsy-related factors as well as properties of the various AEDs. Because treatment is prophylactic and seizures occur at irregular intervals and sometimes have serious consequences, it may be difficult to find the optimal dose on clinical grounds alone. Occasionally, signs of toxicity may be insidious and difficult to interpret, in particular among patients with epilepsy who have associated mental handicaps. The chronic, sometimes lifelong treatment also makes it particularly important to monitor therapy to reduce the risk of long-term adverse effects. These arguments for therapeutic monitoring in epilepsy are valid irrespective of the type of AED. However, the usefulness of drug monitoring also depends on the pharmacologic properties of the drug to be monitored. Therapeutic monitoring is likely to be particularly useful if there is a pronounced interindividual variability in pharmacokinetics, if the kinetics in the individual patient may be altered by drug interactions, concurrent disease, or other conditions, if there is an established correlation between the drug concentration and its therapeutic and toxic effects, and if the therapeutic range is narrow.

The use of drug assays in therapy control rests on the assumption that the relationship between the serum concentration of the active drug and its effects is better than that between dose and effect. Some pharmacologic requirements need to be fulfilled in part or in full to obtain a meaningful relationship between the serum concentration of a drug and its effect. The drug should have a reversible action, and development of tolerance should not occur at receptor sites. It should act by itself and not through metabolites (unless these are measured), and the level of unbound drug at the site of sampling should ideally be equal to the unbound concentration at receptor sites. Hence, although the epilepsy-related rationale for drug


monitoring is similar for all AEDs, the usefulness of this monitoring varies among AEDs, depending on their pharmacological properties.


Most of the older AEDs have a more or less well-defined therapeutic range of serum levels (Table 9.1). This range must not be strictly interpreted, because many of the underlying studies are based on patients with severe epilepsy treated with several AEDs; controlled studies in patients using one drug only for newly diagnosed or moderate epilepsy are scarce. However, most patients are optimally treated with a drug when its steady-state serum levels are maintained within that range (17, 18, 19).

In mild epilepsy, seizure control is often attained at levels lower than the usually recommended range (20, 21, 22). In fact, investigators have suggested that the lower limit of the therapeutic range should be disregarded, and any concentration up to the toxic limit should be considered potentially therapeutic (6). The dose should thus not be increased just to reach the defined therapeutic serum levels in patients who become seizure free with low drug levels. Conversely, some patients with more severe epilepsy need “supratherapeutic” levels to achieve optimal effects. It is also possible that the optimal serum levels differ according to seizure type. Thus, the dose should be titrated to the “optimal serum level” for the individual patient.



Elimination Half-life (h)

Time to Steady-state (days)

Therapeutic Range (µmol/L)

Comment on Value of Therapeutic Range

Older AEDs






Of value






Probably limited use





60-220 nmol/L

Probably limited use






Of value






Of value












Monitor phenobarbital


Valproic acid (sodium valproate)




Of value

Newer AEDsc






Potentially of value






Probably limited use






Potentially of value






Probably limited use






Potentially of value






Potentially of value






Of little value






Potentially of value






Potentially of value

AED, antiepileptic drug; NE, not established.

Concentration dependent.

Monohydroxy derivative.

Therapeutic ranges (µmol/L) are tentative and given in round numbers because of lack of precise documentation.

The therapeutic ranges for phenytoin, phenobarbital, carbamazepine, ethosuximide, valproate, and clonazepam are reasonably well recognized (20,23, 24, 25, 26, 27, 28, 29, 30, 31,32) (Table 9.1). Further studies are necessary to determine the place of serum concentration monitoring for clobazam and for the newer AEDs (Table 9.1). In patients treated with clobazam, the serum levels are in the order of 0.1 to 1.0 µmol/L for the parent drug and 1 to 10 µmol/L for the metabolite, desmethylclobazam (33,34).

A retrospective analysis of patients treated with felbamate reveals that serum concentrations between 210 and 462 µmol/L are associated with optimal seizure control (35). Another study, however, suggests that adverse effects are more frequent in patients with felbamate concentrations >231 µmol/L than with lower concentrations (36).

Wide ranges of gabapentin plasma concentrations have been reported to be associated with seizure control. Therapeutic effects of gabapentin were evident in refractory patients with partial seizures only at serum concentrations >12 µmol/L (37). In another study (38), using high-dose gabapentin in patients with refractory partial seizures, plasma concentrations among responders ranged from 34 to 122 µmol/L.

For lamotrigine, a target range of 4 to 16 µmol/L was suggested early (39), but this range is probably too low because many patients benefit from and tolerate considerably higher concentrations. Morris et al. (40) proposed a target range of 12 to 55 µmol/L based on their own experience and on an open trial of patients with intractable


seizures (41). Most studies have shown a wide range in plasma concentrations associated with seizure control and a considerable overlap in plasma concentrations of responders and nonresponders, as well as between patients with and without side effects (42, 43, 44).

For levetiracetam, the role of therapeutic drug monitoring has not yet been established.

The target range of serum levels for the active metabolite of oxcarbazepine has not yet been well defined. Friis et al. (45) found a mean trough level of approximately 80 µmol/L in their retrospective analysis of 947 patients, but the range in serum concentrations was wide (12 to 160 µmol/L), and the relation to effects and toxicity not analyzed in detail. Similarly, Van Parys and Meinardi (46) reported concentrations ranging from 12 to 128 µmol/L (mean, 70) in seizure-free adult patients. Neither the mean concentration nor the range differed from those in nonresponders. Borusiak et al. (47) reported somewhat higher concentrations in children, 60 to 220 µmol/L (mean, 120). Side effects were more frequent at plasma concentrations of 140 to 160 µmol/L. A population pharmacokinetic-pharmacodynamic assessment was made based on samples from 513 patients participating in three double-blind trials of oxcarbazepine (48). Generally, safety and efficacy responses could adequately be explained by oxcarbazepine dose alone, and plasma concentrations of the metabolite provided limited additional information.

Information on concentration-effect relations with tiagabine is scarce. However, a more pronounced reduction in seizures was observed at trough plasma concentrations >106 µmol/L in a preliminary analysis of data from a clinical trial of patients with complex partial seizures given three different dosages of tiagabine (49).

In contrast to most other AEDs, monitoring of serum levels of vigabatrin is not suitable as a guide to therapy. The mechanism of action of vigabatrin is irreversible inhibition of γ-aminobutyric acid (GABA)-transferase, the enzyme responsible for the catabolism of GABA. Because of the long-lasting inhibition, the antiepileptic effect of vigabatrin long outlasts its presence in serum.

Seizure control has been associated with topiramate serum levels between 10 and 16 µmol/L (50,51). A preliminary report based on topiramate as add-on therapy with other AEDs in patients with refractory epilepsy suggests that levels >12 µmol/L are necessary for effectiveness (52). In monotherapy, patients with topiramate concentrations >30 µmol/L have had a better response than those with lower levels (53).

In add-on studies of zonisamide, favorable clinical responses have been observed at serum levels of 94 to 141 µmol/L (54) and at 33 to 188 µmol/L in children (55). There was, however, a considerable overlap between serum concentrations of responders and nonresponders as well as between serum levels associated with seizure control and those related to side effects (56).

Although clearly further studies are needed, tentative serum concentration ranges for the newer AEDs are given in Table 9.1. As indicated earlier, seizure control is generally reported at a wide range of serum concentrations, and there is a significant overlap with concentrations observed in nonresponders and with patients reporting side effects (6,15). It is clear that the optimal concentration is individual, and this may vary considerably among patients, as indeed is also the case for the older AEDs.

Relationship between Dose and Serum Drug Level


The bioavailability of a drug varies not only with the mode of administration but also between individuals and drugs. Usually, AEDs are given orally, but some can also be given intravenously, intramuscularly, and rectally. After oral administration, maximal serum levels of most AEDs are attained within 2 to 8 hours of the dose intake (6,9). The rate and extent of absorption can be influenced by several factors, as discussed in the following sections.

Biopharmaceutical Formulation.

The rate and extent of absorption are affected by the biopharmaceutical formulation and by differences between nongeneric and generic preparations. The changeover to a preparation with better bioavailability can give serum levels that are too high, with a subsequent risk of drug intoxication. Conversely, a preparation with poor bioavailability can lead to subtherapeutic serum levels and increased seizure frequency (57,58).

Preparation Form.

AEDs are marketed in solutions, syrups, standard tablets, enteric-coated tablets, slow-release tablets, capsules, and suppositories. Standard tablets are absorbed faster than enteric-coated tablets and slow-release (retard) tablets (59).

Gastrointestinal Content.

The degree of absorption of some drugs can be greater when they are taken under fasting conditions, and for other drugs, the converse is true.

Protein Binding and Distribution

On entering the circulation, most AEDs are partly bound to serum proteins and establish an equilibrium between free and bound fractions (6,9,60,61). Only a free drug dissolved in plasma can cross biologic membranes and can interact with specific brain receptors. There is then a dynamic equilibrium between AED molecules in the plasma and in cerebral extracellular fluid. Therefore, the drug concentration in the plasma is assumed to be a measure of drug concentration


in the brain and thus provides a measure of antiepileptic effect because AEDs clearly exert their antiepileptic actions within the brain (62,63). Protein binding differs for each drug, but normally it is quite constant for one AED in the same patient. A reduction in binding can have clinical consequences with highly protein-bound drugs such as phenytoin, valproate, and tiagabine.

Factors Determining the Steady-State Concentration in Serum

The relationship between clinical effect and serum drug levels is evaluated during steady-state conditions. The time lapse before steady-state serum drug levels are attained after initiation of therapy depends on the biologic half-life of the drug (6,9) (Table 9.1). Theoretically, 97% of the final steady-state serum drug level is achieved five half-lives after the initiation of drug therapy. Five times the drug half-life is thus taken as a time interval necessary to achieve steady state. The same rule applies when dosage adjustments are made. Usually, more time is needed because the maintenance dose is often increased gradually. Therefore, the first blood sample related to steady-state concentrations should be drawn about 1 to 2 weeks after initiation of treatment, except for phenobarbital, for which a 3- to 4-week period is recommended because of the longer half-life of this drug. Because carbamazepine increases its own metabolism (autoinduction), the serum drug level may decrease somewhat in the initial dosage period. The time needed before the patient is adjusted to the optimal serum AED level regarding clinical efficacy may, of course, be considerably longer. Generally, the dose should be gradually increased with sufficient time to evaluate the clinical effect to limit adverse reactions.

The rate of elimination of a drug varies from one person to another. Besides genetic factors and interactions resulting from comedication, the rate of elimination depends on age and sex. Fertile women often have faster drug metabolism rates than do men. The rate of metabolism decreases with increasing age for both sexes. When related to age, children tend to require higher drug doses (in milligrams per kilogram) than adults to obtain comparable serum levels. Both renal and liver disease may markedly influence drug protein binding and elimination and can be of clinical importance. Under these conditions, it is therefore especially important that the dose of AED is increased slowly until seizure control is achieved, and the serum concentration is followed and carefully evaluated to avoid intoxications.

Drug Interactions

When a patient is treated with more than one drug, there is often a risk of drug interactions that may result in altered therapeutic outcome. Drug interactions may alter absorption, protein binding, receptor action, metabolism, and excretion of any therapeutic agent (15,51,64,65). The most commonly encountered drug interactions are changes in the rates of biotransformation as a consequence of enzyme induction or inhibition. Interactions between and among AEDs and between these and other drug classes are common. Thus, lower serum AED levels may result in loss in efficacy, and higher levels may result in increased adverse effects.

In long-term treatment of patients with epilepsy, it is important to be aware of the possibility of such drug interactions. Although some drug interactions are not clinically significant, others may cause a marked change in the clinical status of the patient. Change in drug levels in patients treated with the same drugs and dosage varies, and reliable predictions are very difficult to make. Thus, a drug interaction observed in one patient may not be observed in another. When one uses drug combinations that can lead to clinically significant interactions, it is therefore important to be observant for loss of efficacy or clinical signs of intoxication, and it is also necessary to monitor the drug levels 2 to 4 weeks after the addition or withdrawal of a drug. For instance, if an intoxication occurs and the interfering secondary drug cannot be withdrawn, the dose of the primary drug should be reduced, based on appropriate serum drug level monitoring.


Total Serum Drug Levels Versus Free (i.e., Unbound) Serum Drug Levels

Because it is generally assumed that free (i.,e., unbound) drug in plasma is the therapeutically active fraction, the most meaningful way of evaluating clinically significant drug concentrations would be to measure the free drug concentration. However, current analytic methods for quantitation of these drugs measure the total concentration. The drug assays do not distinguish between protein-bound and free drug. Consequently, both fractions are measured and are expressed as a total drug concentration, which routinely is much easier to measure.

An important condition for using the total serum drug concentration as a guideline for the therapeutic or toxic effect is that the protein binding of a certain drug must be constant for the individual patient and the same for all patients. The protein binding differs for each drug depending on its physical and chemical properties, as well as on the physical characteristics of the serum protein. A drug may be either tightly or loosely bound, depending on its affinity for serum proteins. Displacement of a drug from its binding site may result in clinical toxicity because of increased concentration of unbound drug, even though the total level is unchanged.

Normally, the degree of binding is quite constant (9,66). A reduction in binding can take place under special conditions


such as uremia (endogenous displacers). This can be of clinical importance for drugs that are highly bound, such as phenytoin. Newborns with hypoalbuminemia and especially those with hyperbilirubinemia have a reduced drug binding capacity that gives a greater free, and thus therapeutically active, drug fraction in serum. Under these circumstances, a drug with reduced binding capacity and consequently elevated free fraction may be clinically effective or toxic at lower total serum drug levels than would be expected. It is therefore important to consider these aspects when the dosage is determined. Several techniques are available for measurement of the protein binding and the free serum drug levels, including ultrafiltration and equilibrium dialysis (67).

At present, the value of routine monitoring of free drug levels is doubtful. More studies of the relationship between free drug levels and clinical effect are required to evaluate this approach. Therefore, free drug level measurements should be restricted to problem cases. It is of importance in patients who clinically fail to respond to AED therapy because of altered protein binding as a consequence of drug interactions or in special physiologic or pathologic states such as pregnancy, hypoalbuminemia, or hepatic or renal failure (68, 69,70, 71).

Drug Levels in Other Body Fluids

The free drug level can also be derived from the drug concentration in cerebrospinal fluid, saliva, or tears (60,72, 73, 74, 75, 76, 77). Access to cerebrospinal fluid samples is, of course, limited, and studies of the concentration of AEDs in tears have been scarce until recently. Saliva, however, can be an alternative medium to serum for carbamazepine, phenytoin, primidone, and ethosuximide because the concentration in saliva reflects the concentration of free drug in serum. This is not the case for phenobarbital and valproate, owing to the physicochemical properties of these substances (72). However, the concentration of a drug in saliva can be influenced by the sampling conditions and by the contamination of mucus and of unswallowed residual drug and food. Under well-controlled sampling conditions, saliva measurements can be useful for determination of the concentration of free drug, especially when a change in the protein binding is suspected.

Because saliva can be collected by noninvasive techniques, this approach is helpful when multiple serial samples are needed, particularly in children. However, further clinical studies with sensitive analytic methods are needed to evaluate the use of saliva and tears as assay mediums. This is also the case for routine measurements of unbound drug concentration in serum.

Preliminary studies in patients and in healthy persons indicate that valproate concentrations in subcutaneous tissue sampled by microdialysis may reflect unbound serum concentrations (78). Further studies are needed to explore whether this may become a useful method when serial sampling of unbound drug concentrations is required and to ascertain whether the technique could be applied to other AEDs.

Serum Levels of Drug Metabolites

The evaluation of therapeutic drug monitoring will be further refined when the active metabolites of AEDs also can be taken into account. So far, the role of metabolites is not clarified to that extent. One exception is the role of phenobarbital derived from primidone. Primidone is rapidly converted to phenylethylmalonamide and is more slowly converted to phenobarbital. Most physicians routinely prefer to measure and use the concentration of derived phenobarbital during primidone therapy and both primidone and phenobarbital in certain cases only. Routine determination of phenylethylmalonamide is then unnecessary.

Carbamazepine is also a drug with antiepileptic effects resulting from metabolites, the main metabolite being carbamazepine-10,11-epoxide (79). The epoxide levels may range from 10% to 50% of the serum levels of the parent compound. However, there is no constant relationship between the concentration of the epoxide and carbamazepine, and the ratio varies during the day in the individual patient and with comedication (60,80). One should measure both carbamazepine and the epoxide in clinical trials, but at present measuring the metabolite is not indicated for routine monitoring. Oxcarbazepine is the keto derivative of carbamazepine and is rapidly and almost completely metabolized to 10,11-dihydro-10-hydroxycarbazepine, and therefore, the serum level of this metabolite is measured.


The importance of serum drug level monitoring in patients receiving long-term treatment has been emphasized. The clinical conditions for monitoring and determining when the serum drug level ideally should be measured will vary, depending on the pharmacological properties of the AED, but some general guidelines can be issued.

Initiation of Drug Therapy, Dose Adjustment, and Other Medication

Two to 3 weeks after initiation of drug therapy, when steady-state conditions are attained, the serum drug level should be measured and correlated with the clinical effect to see whether the dose is optimal. This is particularly relevant for AEDs with a narrow and well-defined target range. However, a drug level for most AEDs at this point may also be of value as reference for future situations with therapeutic failure. Dosage adjustment is a further indication for


monitoring, in particular of AEDs with dose-dependent kinetics, and it is mandatory after dosage adjustments of phenytoin. The addition of comedication that may cause interactions also prompts monitoring of serum drug levels.

Therapy Failure, Intoxication, and Noncompliance

Inadequate serum drug levels can result from an insufficient dose prescribed (physician), from too low a dose taken (patient compliance) (81), or from a high rate of metabolism (genetic variability or induction). Correspondingly, a high concentration can be caused by too high a dose prescribed or taken or by a low rate of metabolism. Drug compliance is often poor in the treatment of patients with epilepsy.

At high serum drug concentrations, it may be necessary to withdraw the drug for several days. When serum drug concentrations have returned to an appropriate level, therapy can be reinstated at a lower dose.

Other Illnesses

Monitoring is essential in patients with other illnesses and treatment that can influence the disposition of the drug, the water balance, and thus the serum concentration and the seizure control (82). For instance, infections and long-lasting diarrhea may have such an influence.

Monitoring is important in patients with liver and renal dysfunction (83). In patients with severe hepatic disease, drug kinetics may be significantly altered (84, 85, 86). The serum protein binding may be reduced as reported for phenytoin, carbamazepine, and valproate. The same also applies to the capacity of the drug-metabolizing enzymes in the liver. However, the net effect of the total drug clearance may vary considerably, and monitoring of serum levels, together with close clinical observation of the patient, is important.

The clinical picture in renal disease may also vary, and the effect on drug disposition of AEDs may be different in individual patients (84,87). The unbound fraction (not the free concentration) of phenytoin and valproate is increased, but it is rarely necessary to reduce daily doses because of the concomitant increase in total drug clearance. Accordingly, a therapeutic effect may be observed at lower total drug levels than usual. The risk of adverse effects caused by drug interactions is considerably increased in patients with renal or liver disease.


During pregnancy, it is especially important to maintain good seizure control without side effects to avoid harm to the mother and fetus from the seizures and the drugs. Any major changes in AED therapy should be made before the patient conceives. Women considering pregnancy should be placed on the simplest feasible medication regimen. Levels should be closely monitored to determine the lowest dose that will achieve comfortable serum drug levels to keep the risk of drug-induced abnormalities in the child as low as possible.

During pregnancy, several pharmacokinetic parameters of most AEDs are changed significantly, resulting in decreased steady-state serum levels, but the free (i.e., unbound) drug level may be unchanged. Total serum AED levels and, for some AEDs, if possible, free levels, should be measured at regular intervals throughout pregnancy. If an AED dose is increased during pregnancy, the dose should be returned to prepregnancy levels during the first weeks of the puerperium, to avoid toxicity. Drug levels must be checked periodically for at least the first 2 months after delivery (88,89).

Routine Control

Routine control of the serum AED levels is useful because this gives the best basis for comparison if the clinical situation should change. In adults with well-controlled seizures, the serum drug levels may be routinely checked once a year in connection with a medical checkup. Children should be evaluated more frequently during growth when altered drug disposition can be expected. Frequent monitoring is essential at the earliest onset of puberty because metabolic patterns undergo rapid changes at this time. Seizure control can easily worsen because of increasing weight and altered drug disposition.

Sampling Conditions

Whenever possible, the blood sampling time for individual patients should be standardized to ensure comparable conditions. Ideally, the samples should be taken during drug fasting in the morning; in outpatients, the morning dose can be postponed a couple of hours to ensure this condition. For determination of valproate and other AEDs with a short half-life, it is mandatory that the sample be taken during drug fasting in the morning because these drugs have great variations in the serum level between dose intakes. When toxic symptoms of drug are suspected during the day, it is best to draw the sample at the time of maximal serum drug level or at the time when the toxic symptoms are most pronounced. However, one has to keep in mind that the therapeutic ranges are based on trough levels. The following patient data are necessary for a meaningful evaluation of drug levels: age, weight, sex, diagnosis, indications for analysis, clinical conditions of relevance, all drugs in use with total daily doses, and sampling time in relation to the last drug intake.




Numerous methods are available for the determination of serum AED levels, including various gas chromatographic, liquid chromatographic, and immunoassay procedures (2,90). Analytically, it is easier to determine the level of a single drug than that of a mixture of drugs. Because patients often are treated with more than one drug, it is necessary to use selective methods for drug measurement.

An immunoassay for specific measurement of a single drug is often preferred. These assays have several advantages over other currently used methods because they are precise, reproducible, and rapid for determination of drugs in microsamples. However, they are not as suitable for drug screening as are certain chromatographic methods.


Active participation in both internal and external quality control schemes to ensure reliable results is necessary for any laboratory engaged in the routine determination of AEDs. Optimal analytic quality is important for effective therapy; otherwise, the patient can easily be mistreated, and the analytic technique, as well as the laboratory, can become discredited. International cooperation on voluntary quality control schemes of AEDs has improved the analytic performance of many laboratories engaged in therapeutic drug monitoring (91,92).


No strict correlation exists between efficacy and toxicity of AEDs and their serum level for an individual patient. The degree of seizure control and toxicity varies widely among patients with the same serum drug levels. No therapeutic or optimum range is applicable to all patients. The optimum serum drug level depends partly on the severity of the epilepsy and partly on the pharmacodynamic response to a drug. Therefore, drug monitoring is not a substitute for clinical judgment. It is the patient who is being treated, not the serum drug level (3,93, 94, 95, 96).

A serum level judged on a single sample may be misleading for drugs with wide diurnal serum level fluctuations as with carbamazepine and valproate, especially if slow-release tablets are not used. Therefore, the blood sampling time must be standardized whenever possible. If peak-related side effects are suspected, a kinetic profile with several measurements in the same day will provide more information than a single trough level.

One has to keep in mind the analytic problems in determination of AEDs. Accuracy may be a problem in many laboratories in spite of quality control programs. If unexpected values are reported, repeated measurements should be carried out before taking clinical action.


In the future, more emphasis should be placed on rational and cost-effective therapeutic monitoring (97,98). Methods that produce rapid results are of greater relevance and lead to more efficient patient care than assays that involve time lapses of hours or days before the clinician can take action.

For the older AEDs, the therapeutic range and the place of serum drug level monitoring were established long after the introduction of the drugs. Many newer AEDs have been marketed in recent years, and others are in clinical trials (99). The role of therapeutic monitoring of these drugs and of possible comedication needs further investigations. Development and testing of newer AEDs should include an evaluation of therapeutic drug monitoring as early as possible in specifically designed clinical studies, preferably using monotherapy.


  1. Johannessen SI, Morselli PL, Pippenger CE, et al., eds. Antiepileptic therapy: advances in drug monitoring.New York: Raven Press, 1980.
  2. Pippenger CE, Penry JK, Kutt H, eds. Antiepileptic drugs: quantitative analysis and interpretation.New York: Reven Press, 1978.
  3. Chadwick D. Overuse of monitoring of blood concentrations of antiepileptic drugs. BMJ1987;294:723-724.
  4. Schoenenberger RA, Tanasijevic MJ, Jha A, et al. Appropriateness of antiepileptic drug level monitoring. JAMA1995;274: 1622-1626.
  5. Iannuzzi G, Cian P, Fattore C, et al. A multicentre randomized controlled trial on the clinical impact of therapeutic drug monitoring in patients with newly diagnosed epilepsy.Epilepsia2000; 41:222-230.
  6. Perucca E. Is there a role for therapeutic drug monitoring of the new anticonvulsants? Clin Pharmacokinet2000;38:191-204.
  7. Commission on Antiepileptic Drugs. Guidelines for therapeutic monitoring on antiepileptic drugs. Epilepsia1993;34:585-587.
  8. Levy RH, Pitlick WH, Eichelbaum M, et al., eds. Metabolism of antiepileptic drugs.New York: Raven Press, 1984.
  9. Morselli PL, Franco-Morselli R. Clinical pharmacokinetics of antiepileptic drugs in adults. Pharmacol Ther1980;10:65-101.
  10. Schmidt D, Jacob R. Clinical and laboratory monitoring of antiepileptic medication. In: Wyllie E, ed. The treatment of epilepsy: principles and practice.Philadelphia: Lea & Febiger, 1993:798-809.
  11. Johannessen SI. Plasma drug concentration monitoring of anticonvulsants: practical guidelines. CNS Drugs1997;7:349-365.
  12. Eadie MJ. Therapeutic drug monitoring: antiepileptic drugs. Br J Clin Pharmacol1998;46:185-193.
  13. Brodie MJ. Routine measurement of new antiepileptic drug concentrations: a critique and a prediction. In: French J, Leppik I, Dichter MA, eds. Antiepileptic drug development. Adv Neurol1998;76:223.



  1. Patsalos PN. New antiepileptic drugs. Ann Clin Biochem1999; 36:10-19.
  2. Tomson T, Johannessen SI. Therapeutic monitoring of the new antiepileptic drugs. Eur J Clin Pharmacol2000;55:697-705.
  3. Glauser TA, Pippenger CE. Controversies in blood-level monitoring: re-examining its role in the treatment of epilepsy. Epilepsia2000;41[Suppl 8]:S6-S15.
  4. Schmidt D, Einicke I, Haenel F. The influence of seizure type on the efficacy of plasma concentrations of phenytoin, phenobarbital, and carbamazepine. Arch Neurol1986;43:263-265.
  5. Turnbull DM, Rawlins MD, Weightman D, et al. “Therapeutic” serum concentration of phenytoin: the influence of seizure type. J Neurol Neurosurg Psychiatry1984;47:231-234.
  6. Vaida FJE, Aicardi J. Reassessment of the concept of a therapeutic range of anticonvulsant plasma levels. Dev Med Child Neurol1983;25:660-671.
  7. Lund L. Anticonvulsant effect of diphenylhydantoin relative to plasma levels. Arch Neurol1974;31:289-294.
  8. Shorvon SD, Galbraith AW, Laundy M, et al., eds. Antiepileptic therapy: advances in drug monitoring.New York: Raven Press, 1980:213-220.
  9. Strandjord RE, Johannessen SI. Carbamazepine as the only drug in patients with epilepsy: serum levels and clinical effect. In: Johannessen SI, Morselli PL, Pippenger CE, et al., eds. Antiepileptic therapy: advances in drug monitoring.New York: Raven Press, 1980:229-235.
  10. Baruzzi A, Bordo B, Bossi L, et al. Plasma levels of dipropylacetate and clonazepam in epileptic patients. Int J Clin Pharmacol1977;15:403-408.
  11. Cereghino JJ. Serum carbamazepine concentration and clinical control. Adv Neurol1975;11:309-329.
  12. Eichelbaum M, Bertilsson L, Lund L, et al. Plasma levels of carbamazepine and carbamazepine 10,11-epoxide during treatment of epilepsy. Eur J Clin Pharmacol1976;9:417-421.
  13. Gram L, Flachs H, Würtz-Jørgensen A, et al. Sodium valproate, serum level and clinical effect in epilepsy: a controlled study. Epilepsia1979;20:303-312.
  14. Henriksen O, Johannessen SI. Clinical observations of sodium valproate in children: an evaluation of therapeutic serum levels. In: Johannessen SI. Morselli PL, Pippenger CE, et al., eds. Antiepileptic therapy: advances in drug monitoring.New York: Raven Press, 1980:253-261.
  15. Monaco F, Riccio A, Benna P, et al. Further observations on carbamazepine plasma levels in epileptic patients: relationships with therapeutic and side effects. Neurology1976;26:936-943.
  16. Penry JK, Porter RJ, Dreifuss FE. Ethosuximide: relation of plasma levels to clinical control. In: Woodbury DM, Penry JK, Schmidt RP, eds. Antiepileptic drugs.New York: Raven Press, 1972:431-441.
  17. Schottelius DD, Fincham RW. Clinical application of serum primidone levels. In: Pippenger CE, Penry JK, Kutt H, eds. Antiepileptic drugs: quantitative analysis and interpretation.New York: Raven Press, 1978:273-282.
  18. Sjö O, Hvidberg EF, Naestoft J, et al. Pharmacokinetics and side-effects of clonazepam and its 7-amino-metabolite in man. Eur J Clin Pharmacol1975;8:249-254.
  19. Sundqvist A, Tomson T, Lundkvist B. Valproic acid in patients with juvenile myoclonic epilepsy on monotherapy: a dose-effect study. Ther Drug Monit1998;20:149-157.
  20. Streete JM, Berry DJ, Newbery JE. The analysis of dobazam and its metabolite desmethylclobazam by high-performance liquid chromatography. Ther Drug Monit1991; 13:339-344.
  21. Wang J, Hug D, Gautschi K, et al. Clobazam for treatment of epilepsy. J Epilepsy1993;6:180-184.
  22. Troupin AS, Montouris G, Hussein G. Felbamate: therapeutic range and other kinetic information. J Epilepsy1997;10:26-31.
  23. Harden CL, Trifiletti T, Kutt H. Felbamate levels in patients with epilepsy. Epilepsia1996;37:280-283.
  24. Sivenius J, Kälviäinen R, Ylinen A, et al. Double blind study of gabapentin in the treatment of partial seizures. Epilepsia1991; 32:539-542.
  25. Wilson EA, Sills GJ, Forrest G, et al. High dose gabapentin in refractory partial epilepsy: clinical observations in 50 patients. Epilepsy Res1998;29:161-166.
  26. Brodie MJ, Richens A, Yuen AWC. Double-blind comparison of lamotrigine and carbamazepine in newly diagnosed epilepsy. Lancet1995;345:476-479.
  27. Morris RG, Black AB, Harris AL, et al. Lamotrigine and therapeutic drug monitoring: retrospective survey following the introduction of a routine service. Br J Clin Pharmacol1998;46: 547-551.
  28. Schapel G, Black A, Lam E, et al. Combination vigabatrin and lamotrigine therapy for intractable epilepsy. Seizure1996:5: 51-56.
  29. Bartoli A, Guerrini R, Belmonte A, et al. The influence of dosage, age and comedication on steady-state plasma lamotrigine concentrations in epileptic children: a prospective study with preliminary assessment of correlations with clinical response. Ther Drug Monit1997:19:252-260.
  30. Kilpatric ES, Forrest G, Brodie M. Concentration-effect and concentration-toxicity relations with lamotrigine: a prospective study. Epilepsia1996;37:534-538.
  31. Eriksson A-S, Nergårdh A, Hoppu K. The efficacy of lamotrigine in children and adolescents with refractory generalized epilepsy: a randomized, double-blind, cross-over study.Epilepsia1998;39:495-501.
  32. Friis ML, Kristensen O, Boas J, et al. Therapeutic experiences with 947 epileptic out-patients in oxcarbazepine treatment. Acta Neurol Scand1993;87:224-227.
  33. Van Parys JAP, Meinardi H. Survey of 260 patients treated with oxcarbazepine (Trileptal) on a named-patient basis. Epilepsy Res1994;19:79-85.
  34. Borusiak P, Korn-Merker E, Holert N, et al. Oxcarbazepine in treatment of childhood epilepsy: a survey of 46 children and adolescents. J Epilepsy1998;11:355-360.
  35. Nedelman JR, Gasparini M, Hossain M, et al. Oxcarbazepine: analysis of concentration-efficacy/safety relationships. Neurology1999;52[Suppl 2]:A524-A525.
  36. Rowan AJ, Gustavson L, Shu V, et al. Dose concentration relationship in a multicentre tiagabine (Gabitril) trial. Epilepsia1997;38[Suppl 3]:40.
  37. Reife RA, Pledger G, Doose D, et al. Topiramate PK/PD analysis. Epilepsia1995;36[Suppl 3]:S152.
  38. Perruca E, Bialer M. The clinical pharmacokinetics of the newer antiepileptic drugs: focus on topiramate, zonisamide and tiagabine. Clin Pharmacokinet1996;31:29-46.
  39. Penovich PE, Schroeder-Gustafson M, Gates JR, et al. Clinical experience with topiramate: correlation of serum levels with efficacy and adverse events. Epilepsia1997;38[Suppl 8]:181.
  40. Twyman RE, Ben-Menachem E, Veloso F, et al. Plasma topiramate (TPM) concentrations vs therapeutic response during monotherapy. Epilepsia1999;40[Suppl 7]:111-112.
  41. Wilensky AJ, Friel PN, Ojemann LM, et al. Zonisamide in epilepsy: a pilot study. Epilepsia1985;26:212-220.
  42. Mimaki T, Mino M, Sugimoto T, et al. Antiepileptic effect and serum levels of zonisamide in epileptic patients with refractory seizures. In: Sunshine I, ed. Recent developments in therapeutic drug monitoring and clinical toxicology.Marcel Dekker, New York, 1992:437-442.
  43. Mimaki T. Clinical pharmacology and therapeutic drug monitoring of zonisamide. Ther Drug Monit1998;20:593-597.



  1. Lund L. Clinical significance of generic inequivalence of two different pharmaceutical formulations of phenytoin. Eur J Clin Pharmacol1974;7:119-124.
  2. Peterson H de Coudres. Brand-name antiepileptic drugs versus generics. In: Resor SR Jr, Kutt H, eds. The medical treatment of epilepsy.New York: Marcel Dekker, 1992:493-495.
  3. Johannessen SI, Henriksen O. Comparison of the serum concentration profiles of Tegretol and two new slow-release preparations. In: Wolf P, Dam M, Janz D, et al., eds.Advances in epileptology: XVth epilepsy international symposium.New York: Raven Press, 1987:421-424.
  4. Johannessen SI, Gerna M, Bakke J, et al. CSF concentrations and serum protein binding of carbamazepine and carbamazepine10,11-epoxide in epileptic patients. Br J Clin Pharmacol1976;3: 575-582.
  5. Johannessen SI, Strandjord RE. Absorption and protein binding in serum of several anti-epileptic drugs. In: Schneider H, Janz D, Gardner-Thorpe C, et al., eds. Clinical pharmacology of antiepileptic drugs.New York: Springer-Verlag, 1975:262-273.
  6. Morselli PL, Baruzzi A, Gerna M, et al. Carbamazepine and carbamazepine-10,11-epoxide concentrations in human brain. Br J Clin Pharmacol1977;4:535-540.
  7. Sherwin AL, Eisen AA, Sokolowski CD. Anticonvulsant drugs in human epileptogenic brain: correlation of phenobarbital and diphenylhydantoin levels with plasma. Arch Neurol1973;29:73-77.
  8. Kutt H. Pharmacokinetic interactions with antiepileptic medication. In: Wyllie E, ed. The treatment of epilepsy. Principles and practice. Philadelphia: Lea & Febiger, 1993;775-784.
  9. Pitlick WH, ed. Antiepileptic drug interactions. New York: Demos Publications, 1989.
  10. Johannessen SI. Antiepileptic drugs: pharmacokinetic and clinical aspects. Ther Drug Monit1981;3:17-37.
  11. Pacifici GM, Viani A. Methods of determining plasma and tissue binding of drug: pharmacokinetic consequences. Clin Pharmacokinet1992;23:449-468.
  12. Baird-Lambert J, Manglick MP, Wall M, et al. Identifying patients who might benefit from free phenytoin monitoring. Ther Drug Monit1987;9:134-138.
  13. Lenn NJ, Robertson M. Clinical utility of unbound antiepileptic drug blood levels in the management of epilepsy. Neurology1992;42:988-990.
  14. Levy RH, Schmidt D. Utility of free level monitoring of antiepileptic drugs. Epilepsia1985;25:199-205.
  15. Svensson CK, Woodruff MN, Baxter JG, et al. Free drug concentration monitoring in clinical practice: rationale and current status. Clin Pharmacokinet1986;11:450-469.
  16. Blom GF, Guelen PJM. The distribution of anti-epileptic drugs between serum, saliva and cerebrospinal fluid. In: Gardner-Thorpe C, Janz D, Meinardi H, et al., eds.Antiepileptic drug monitoring.Tunbridge Wells, UK: Pitman Medical, 1977:287-297.
  17. Drobitch RK, Svensson CK. Therapeutic drug monitoring in saliva: an update. Clin Pharmacokinet1992;23:365-379.
  18. Johannessen SI, Henriksen O. Serum levels of di-n-propylacetate in epileptic patients. Pharm Weekbl [Sci]1977;112:287-289.
  19. Johannessen SI, Strandjord RE. Concentration of carbamazepine (Tegretol) in serum and cerebrospinal fluid in patients with epilepsy. Epilepsia1973;14:373-379.
  20. Monaco F, Piredda S, Mastropaolo C, et al. Diphenylhydantoin and primidone in tears. Epilepsia1981;22:185-188.
  21. Schmidt D, Kupferberg HJ. Diphenylhydantoin, phenobarbital, and primidone in saliva, plasma, and cerebrospinal fluid. Epilepsia1975;16:735-741.
  22. Lindberger M, Tomson T, Ståhle L. Validation of microdialysis sampling for subcutaneous extracellular valproic acid in humans. Ther Drug Monitor1998;20:358-362.
  23. Tomson T, Almqvist O, Nilsson BY, et al. Carbamazepine-epoxide in epilepsy: a pilot study. Arch Neurol1990;47:888-892.
  24. Johannessen SI, Baruzzi A, Gomeni R, et al. Further observations on carbamazepine and carbamazepine-10,11-epoxide kinetics in epileptic patients. In: Gardner-Thorpe C, Janz D, Meinardi H, et al., eds. Antiepileptic drug monitoring.Tunbridge Wells, UK: Pitman Medical, 1977:110-127.
  25. Leppik IE. Compliance in the treatment of epilepsy. In: Wyllie E, ed. The treatment of epilepsy: principles and practice.Philadelphia: Lea & Febiger, 1993:810-816.
  26. Kutt H. Effect of acute and chronic diseases on the disposition of antiepileptic drugs. In: Morselli PL, Pippenger CE, Penry JK, eds. Antiepileptic drug therapy in pediatrics.New York: Raven Press, 1983:293-302.
  27. Asconapé JJ, Penry JK. Use of antiepileptic drugs in the presence of liver and kidney diseases: a review. Epilepsia1982;23[Suppl 1]:S65-S79.
  28. Hooper WD, Bochner F, Eadie MJ, et al. Plasma protein binding of diphenylhydantoin: effects of sex hormones, renal and hepatic diseases. Clin Pharmacol Ther1974; 15:276-282.
  29. Hooper WD, Dubetz DK, Bochner F, et al. Plasma protein binding of carbamazepine. Clin Pharmacol Ther1975;17:433-440.
  30. Klotz U, Rapp T, Müller WA. Disposition of valproic acid in patients with liver disease. Eur J Clin Pharmacol1978;13:55-60.
  31. Gugler R, Müller G. Plasma protein binding of valproic acid in healthy subjects and in patients with renal disease. Br J Clin Pharmacol1978;5:441-446.
  32. Tomson T, Gram L, Sillanpää M, et al. Recommendation for the management and care of pregnant women with epilepsy. In: Tomson T, Gram L, Sillanpää M, et al., eds.Epilepsy and pregnancy.Petersfield, XX: Wrightson Biomedical Publishing, 1997:201-208.
  33. Johannessen SI. Pharmacokinetics of valproate in pregnancy: mother, foetus, newborn. Pharm Weekbl [Sci]1992;14:114-117.
  34. Meijer JWA. Knowledge, attitude, and practice in antiepileptic drug monitoring. Acta Neurol Scand Suppl1991;83:134.
  35. Pippenger CE, Penry JK, White BG, et al. Interlaboratory variability in determination of plasma antiepileptic drug concentrations. Arch Neurol1976;33:351-355.
  36. Wilson JF, Tsanaclis LM, Perrett JE, et al. Performance of techniques for measurement of therapeutic drugs in serum: a comparison based on external quality assessment data. Ther Drug Monit1992;14:98-106.
  37. Beardsley RS, Freeman JM, Appel FA. Anticonvulsant serum levels are useful only if the physician appropriately uses them: an assessment of the impact of providing serum level data to the physicians. Epilepsia1983;24:330-335.
  38. Dodson EW. Level off. Neurology1989;39:1009-1010.
  39. Pena AIA, Lope ES. Can a single measurement of carbamazepine suffice for therapeutic monitoring? Clin Chem1987;33:812-813.
  40. Woo E, Chan YM, Yu YL, et al. If a well-stabilized epileptic patient has a subtherapeutic antiepileptic drug level, should the dose be increased? A randomized prospective study. Epilepsia1988;29:129-139.
  41. Pippenger CE. The cost-effectiveness of therapeutic drug monitoring. Ther Drug Monit1990; 12:418.
  42. Vozeh S. Cost-effectiveness of therapeutic drug monitoring. Clin Pharmacokinet1987;13:131-140.
  43. Bialer M, Johannessen SI, Levy RH, et al. Progress report on new antiepileptic drugs: a summary of the Fifth Eilat Conference. Epilepsy Res2001;43:11-58.