Saliva Aed Tdm 2013

REVIEW ARTICLE

Therapeutic Drug Monitoring of AntiepilepticDrugs by Use of Saliva

Philip N. Patsalos, FRCPath, PhD* and Dave J. Berry, FRCPath, PhD

Abstract: Blood (serum/plasma) antiepileptic drug (AED) therapeu-tic drug monitoring (TDM) has proven to be an invaluable surrogatemarker for individualizing and optimizing the drug management ofpatients with epilepsy. Since 1989, there has been an exponentialincrease in AEDs with 23 currently licensed for clinical use, andrecently, there has been renewed and extensive interest in the use ofsaliva as an alternative matrix for AED TDM. The advantages of salivainclude the fact that for many AEDs it reects the free (pharmacolog-ically active) concentration in serum; it is readily sampled, can besampled repetitively, and sampling is noninvasive; does not require theexpertise of a phlebotomist; and is preferred by many patients,particularly children and the elderly. For each AED, this reviewsummarizes the key pharmacokinetic characteristics relevant to thepractice of TDM, discusses the use of other biological matrices withparticular emphasis on saliva and the evidence that saliva concentrationreects those in serum. Also discussed are the indications for salivaryAED TDM, the key factors to consider when saliva sampling is to beundertaken, and nally, a practical protocol is described so as to enableAED TDM to be applied optimally and effectively in the clinicalsetting. Overall, there is compelling evidence that salivary TDM can beusefully applied so as to optimize the treatment of epilepsy withcarbamazepine, clobazam, ethosuximide, gabapentin, lacosamide,lamotrigine, levetiracetam, oxcarbazepine, phenobarbital, phenytoin,primidone, topiramate, and zonisamide. Salivary TDM of valproic acidis probably not helpful, whereas for clonazepam, eslicarbazepineacetate, felbamate, pregabalin, retigabine, runamide, stiripentol,tiagabine, and vigabatrin, the data are sparse or nonexistent.

Key Words: saliva, antiepileptic drugs, therapeutic drug monitoring,practical protocol for saliva antiepileptic drug monitoring

(Ther Drug Monit 2013;35:429)

INTRODUCTIONMeasuring antiepileptic drugs (AEDs) in serum or

plasma as an aid to personalizing drug therapy is now a well-established practice in the treatment of epilepsy, and guidelinesare published that indicate the particular features of epilepsy andthe properties of AEDs that make the practice so benecial.1

The goal of AED therapeutic drug monitoring (TDM) is tooptimize a patients clinical outcome by supporting the man-agement of their medication regimen with the assistance ofmeasured drug concentrations/levels. The reason why TDMhas emerged as an important adjunct to treatment with theAEDs arises from the fact that for an individual patient identi-fying the optimal dose on clinical grounds alone can be difcultand there are many reasons for this including the following: (1)AED treatment is prophylactic and, because seizures occur atirregular intervals, it is often difcult to ascertain whether theprescribed dose will be sufcient to produce long-term seizurecontrol; (2) clinical symptoms and signs of toxicity are notalways readily detectable; (3) the correlation between AEDserum concentration and the clinical effects is much better thanthat between the dose and effect; and (4) there are no directlaboratory markers for clinical efcacy or AED toxicity.

Although reasonably well-dened reference ranges (targetranges) have been established for most of the AEDs,14 one sizedoes not t all, and individual differences in the nature andseverity of epilepsy result in the effective, nontoxic AED con-centration being extremely variable; seizures in some patientscan be well managed at serum concentrations below the targetrange, whereas other patients need and tolerate drug concen-tration in excess of the range.1,2 Furthermore, many factorscause unpredictable and sometimes large differences betweenindividuals in pharmacokinetics and disposition of AEDs,which makes it impossible to predict the optimum dose fora particular patient and measuring a serum concentration willoften be the most effective way to guide treatment. Indeed, theconcept of the individual therapeutic range has been champ-ioned as the ideal practice parameter for bespoke AED ther-apy,1 and a similar approach has recently been advocated forpsychiatric drug therapy.5

Although AED TDM for the treatment of epilepsy wasinitially developed and validated for the few drugs that wereavailable during the 1960s1980s, a further 17 drugs have beenintroduced since 1989 some of which are also effective formanaging other neurological disorders (Table 1). The clinicaltrials of investigational AEDs are undertaken primarily toestablish safety, ascertain pharmacokinetics, and dosagerange, their drugdrug interaction proles, their efcacy overplacebo, and to identify acute adverse effects.6 These are the

Received for publication May 24, 2012; accepted October 19, 2012.From the *Pharmacology and Therapeutics Unit, Department of Clinical and

Experimental Epilepsy, UCL-Institute of Neurology, London, UnitedKingdom; and Epilepsy Society, Chalfont Centre for Epilepsy, ChalfontSt Peter, United Kingdom.

The work undertaken by Professor P. N. Patsalos was performed at UCLH/UCL and received a proportion of funding from the Department ofHealths NIHR Biomedical Research Centres funding scheme. ProfessorPatsalos has received during the past year speakers or consultancy feesfrom the following pharmaceutical companies: Eisai, Sano Aventis, andUCB Pharma.

Correspondence: Philip N. Patsalos, FRCPath, PhD, Department of Clinicaland Experimental Epilepsy, UCL-Institute of Neurology, Queen Square,London WC1N 3BG, United Kingdom (e-mail: [email protected]).

Copyright 2013 by Lippincott Williams & Wilkins

4 Ther Drug Monit Volume 35, Number 1, February 2013

characteristics that must be documented to achieve regula-tory approval. Although serum concentration measurementsof the investigational AEDs are undertaken (often retrospec-tively) during the clinical trial process, informationregarding the serum concentration to effect/toxicity interre-lationship is rarely evaluated at this time. Although therange of serum concentrations determined at the dose rangesinvestigated during clinical trials of a new AED give someuseful information regarding a putative reference range, thecorrelation with clinical effect is rarely evaluated. Neverthe-less, this information can prove useful clinically, particu-larly when it is remembered that serum concentrationmeasurements should be used in the context of the patientsclinical presentation (ie, treat the patient not the serum con-centration). The indications for AED TDM are shown inTable 2.

The aim of this review is to discuss the potential useof saliva as a matrix to undertake AED TDM. First, theadvantages and disadvantages of using various biologicalmatrices with particular emphasis on saliva will bereviewed. Second, for each AED, the key pharmacokineticcharacteristics relevant to the practice of TDM are pre-sented along with the evidence that saliva concentrationsreect those in serum. Third, indications for salivary AEDTDM are emphasized along with the key factors to considerwhen saliva sampling is to be undertaken. Lastly, andnally, a practical protocol is described so as to enable

AEDS TDM to be applied optimally and effectively in theclinical setting. Search strategy and selection criteria: Thisreview is based on published articles and searches inPubMed and Google Scholar up to April 2012, in additionto references from relevant articles. Primary sources werepreferred, but abstracts are included where no subsequentpeer reviewed article was published. Review articles ofimportance were also used. The search terms included thevarious AEDs: carbamazepine, clobazam, clonazepam,eslicarbazepine acetate, ethosuximide, felbamate, gabapen-tin, lacosamide, lamotrigine, levetiracetam, oxcarbazepine,phenobarbital, phenytoin, pregabalin, primidone, retiga-bine, runamide, stiripentol, tiagabine, topiramate, valproicacid, vigabatrin, and zonisamide. Also, the terms saliva,hair, cerebrospinal uid (CSF), tears, dried spot blood, andTDM were searched.

TABLE 2. Indications for AED Therapeutic Drug Monitoring

Indication Comment

1 After initialization of AEDtreatment or after doseadjustment

This allows the pursuance ofa preselected reference range forthe individual patient.

2 Upon achievement of optimumdesired clinical response

Seizure freedom is the optimumoutcome, but for many patients,optimum seizure control withminimal adverse effects ismore readily achieved. Theindividual therapeutic rangecan be established.

3 To determine the magnitude ofa dose change

This is particularly importantfor AEDs that show dose-dependent pharmacokinetics(eg, phenytoin, carbamazepine,valproate, gabapentin,stiripentol, and runamide).

4 When toxicity is difcult todifferentially diagnose or whentoxicity is difcult to assessclinically

Concentration-related AEDtoxicity is more readilyidentied and is particularlyhelpful when young childrenor patients with mentaldisability are being evaluated.

5 When seizures persist despite theprescribing of an adequate/typical dosage

This may identify a fastmetabolizer or a patient thatis noncomplying with theirAED medication.

6 When pharmacokineticvariability is expected

This is a signicant category ofpatients and includes children,the elderly, during pregnancy,hepatic disease, renal disease,various pathologies,postsurgery, and drugdruginteractions.

7 When a formulation change hasoccurred

This includes brand-to-genericand generic-to-generic switches.

8 The clinical response hasunexpectantly changed

The cause of the change couldbe readily identied as itcould be the consequenceof many reasons.

9 Poor compliance suspected Recent noncompliance can bereadily identied. However,long-term compliance orvariable compliance cannotbe identied.

TABLE 1. Introduction of Antiepileptic Drugs in the UnitedKingdom*

Drug Year of Introduction

Phenobarbital 1912

Phenytoin 1938

Primidone 1952

Ethosuximide 1960

Carbamazepine 1963

Valproate 1974

Clonazepam 1974

Clobazam 1982

Vigabatrin 1989

Lamotrigine 1991

Gabapentin 1993

Felbamate 1993

Topiramate 1995

Tiagabine 1998

Oxcarbazepine 2000

Levetiracetam 2000

Pregabalin 2004

Zonisamide 2005

Runamide 2007

Stiripentol 2007

Lacosamide 2008

Eslicarbazepine acetate 2009

Retigabine 2011

*Although in general the order of drug introduction is similar in Europe and theUnited States.

Ther Drug Monit Volume 35, Number 1, February 2013 Saliva Monitoring of Antiepileptic Drugs

2013 Lippincott Williams & Wilkins 5

BIOLOGICAL MATRIXESA variety of biological matrixes have been employed

for TDM of AEDs, and these are reviewed in this section. Thevarious matrices have not been validated for all drugs; inparticular, some have not been investigated with respect to thenewer AEDs. When the nal analytical step is chromatographic,any of the matrices requires some form of sample pretreatment/extraction before the nal analytical process. If immunoassaysare used, theymust also be matrix validated; furthermore, immu-noassays are not commercially available for all AEDs discussedin this review.

BloodTDM is normally undertaken in serum or plasma, and

these matrixes can be used interchangeably because no differ-ences in AED concentration have been demonstrated betweenthem. Many AEDs are bound to serum proteins, and allanalytical methods measure the total serum/plasma concentra-tion [ie, the sum of drug bound to serum protein (usuallyalbumin) and free nonprotein bound drug]. Although thisparameter will sufce in most clinical settings, it is only thefree nonprotein bound drug that can cross the bloodbrainbarrier to equilibrate with brain receptors and produce thepharmacological effect. Therefore, one should ideally measurethe free nonprotein bound drug concentration, and there areclinical situations where protein binding is disturbed and mon-itoring the free nonprotein bound (pharmacologically effec-tive) drug would be more appropriate. This is particularly truefor those AEDs that are extensively protein bound (eg, phenyt-oin, carbamazepine, and valproate). Settings where proteinbinding can be impaired include the following: Hypoalbuminemia: This can occur with renal disease,liver disease, pregnancy, old age, postsurgically, and arange of other pathological conditions (see Pathologicalstates).

Conditions in which endogenous protein displacing agentsaccumulate, for example, uremia.

Binding displacement by another signicantly protein-bound drug.The extent to which a binding disturbance occurs is

unpredictable and may differ from patient to patient and alsowithin the same patient over time. If the free drug concentrationincreases, the determination of total serum concentration willprovide an underestimate of the quantity of nonprotein bounddrug with therapeutic and toxic effects being observed at totalconcentrations that are lower than expected.

The 3 most common methodologies that are used toseparate the free from the bound components in serum areequilibrium dialysis, ultraltration, and ultracentrifugation, withultracentrifugation being the method of choice in pathology lab-oratories.79 However, the ultracentrifugation method is time con-suming and labor intensive to perform and therefore more costlyand thus saliva, which is a natural ultraltrate of serum, is a supe-rior matrix for the measurement of free drug concentrations.

Dried Blood SpotDried blood spot applications in TDM were recently

reviewed by Edelbroek et al,10 and the approach has been

used for a variety of AEDs including carbamazepine, phenytoin,phenobarbital, oxcarbazepine, lamotrigine, runamide, gaba-pentin, and topiramate.1116 The specimens can be collectedby a nger prick at an agreed time either by the patients them-selves or by a carer. After pricking the nger, a blood spot ofadequate size (usually 10 mm) is placed onto an absorbentpaper where it dries. The sample is then sent together withan AED sample request form to the laboratory (this can beposted). Upon receipt, the laboratory can punch a standard sizedisc from the paper (eg, an 8-mm disk would contain approx-imately 15 mL of blood), and the dried blood spot is thenextracted/processed before chromatographic examination.

Advantages of the dried blood spot include the fact thatsampling is simple, safe, and can be carried out in a homeenvironment without the need of a trained phlebotomist. Onlya small amount of blood is required, and if necessary, a sequenceof specimens can be easily collected in 1 day. Specimens canalways be collected at the optimum time, and the result can beavailable for a subsequent clinic visit thus allowing immediateadjustment of medication. Thus, the technique is patient friendlyand has a better cost benet than phlebotomy. Dried blood spotsamples are particularly helpful for patients that are difcult tobleed and also when a specimen needs collecting at a particulartime so as to ascertain whether or not symptoms (eg, transienttoxicity) are drug related.

There are also some disadvantages, for example, thequality of both the blood spot and paper affect the result andsome blood spots are not suitable for testing; also, some patientsare resistant to nger prick.

TearsSome AEDs (eg, phenobarbital, carbamazepine, phenyt-

oin, primidone, ethosuximide, and valproic acid) are transportedinto tears, and the concentration represents the free nonproteinbound concentration of the drug in serum.1722 Furthermore, tearuid, especially after stimulation, is more homogeneous andmore constant in its composition compared with saliva. How-ever, collection of tear samples is rather cumbersome as itinvolves the use of a capillary tube and is considered by manypatients to be invasive because they cannot readily producetears particularly the quantity (100 mL) currently required fordrug analysis. Although it is stated that lacrimation is frequentin children, in adults, brisk tearing has been provoked by smok-ing and/or the snifng of formaldehydethe latter methodsbeing rather undesirable.

SweatSome AEDs are secreted into sweat and Parnas et al23

reported that phenytoin, carbamazepine, and phenobarbital wereall present in sweat and that phenytoin concentrations corre-sponded to the free nonprotein bound fraction in serum andwere independent of sweat ow. However, phenobarbital sweatconcentration increased with sweat ow. Phenobarbital has alsobeen determined in sweat collected into a sweat patch where itspresence is easily demonstrated but this procedure would onlybe useful for documenting drug use over the period duringwhich the patch was applied.24 At present, sweat isnot a very useful matrix for TDM because of the practicalities

Patsalos and Berry Ther Drug Monit Volume 35, Number 1, February 2013

6 2013 Lippincott Williams & Wilkins

associated with owing sweat collection and/or the interpre-tation of the concentration if collected into a patch.

Cerebrospinal FluidMeasurement of CSF concentrations of neuroactive drugs,

including AEDs, is important because such concentrations areconsidered to reect those occurring in the brain, which resultin the pharmacological effect of the drug (eg, anticonvulsant oradverse effects). Furthermore, CSF concentrations are consid-ered to reect the free, nonprotein bound serum concentration.

The transfer of the rst-generation AEDs (carbamazepine,phenytoin, phenobarbital, and valproate) into CSF has been wellstudied, and the CSF concentration generally reects their freenonprotein bound concentration in serum.20,21,2528 Morerecently, some of the second-generation drugs, for example,gabapentin, oxcarbazepine, lamotrigine, levetiracetam, vigabatrin,and topiramate2933 and third-generation AEDs, for example, esli-carbazepine acetate34 have been investigated.

Because for many AEDs the CSF concentration reectsthe free nonprotein bound drug concentration, this would prob-ably be a meaningful matrix for TDM purposes; however, forsome AEDs CSF does not reect free serum concentration forexample, gabapentin and pregabalin.3033 This lack of correlationis probably due to the mechanism by which gabapentin andpregabalin are distributed throughout the body, that is, theL-amino acid transporter, which is saturable so that transportationdoes not occur linearly. In the case of pregabalin, the CSF/serumarea under the concentration versus time curve 024 hours ratiowas 0.098 6 0.016.33 The invasive procedure (ie, lumber punc-ture) needed to collect a CSF samples negates the practical use ofCSF for AED TDM.

HairThe root of every growing hair is constantly exposed to

any drug that is circulating in the blood. The drug(s) are thussequestered into the hair structure and because head hair growsat approximately 1 cm/mo, if only a single drug exposureoccurred the portion of hair containing the drug would emergefrom the scalp after 6/7 days. This small section would thengrow away from the scalp, and by sampling the hair in 1-cmlengths, it is possible to assay the drug concentration in the hairsections and pinpoint the time of exposure. However, in patientsprescribed maintenance drug treatment, provided that they areadherent, the hair root will be exposed to a constant, steady-statedrug concentration in blood and this would be reected by aconstant concentration of the drug along the hair shaft. Hair canthus be used to record the history of drug exposure and ascertainvariable and intermittent compliance. The AEDs that have beenreported to be transported into hair include carbamazepine,phenytoin, valproic acid, oxcarbazepine, and lamotrigine.3539

Although hair analysis would not, therefore be helpfulfor day to day TDM, these principles have been applied todemonstrate compliance with carbamazepine and oxcarbaze-pine treatment in adult patients,3941 compliance of carbamaze-pine and lamotrigine during pregnancy38 and to differentiatebetween chronic and acute carbamazepine intoxication.42

The disadvantages of hair analyses for AED TDM includethe fact that many factors impact on the amount of drugdeposited into hair, for example, melanin content, and whether

cosmetic hair treatments, for example, bleaching, dying, etc,remove drugs that are bound into the hair structure and thus causeinterpatient concentrations to vary signicantly. Also, becausedrugs are incorporated into the hair structure, the analyticalprocess requires an initial digestion to release them before thechosen method of detection and quantication. Finally, althoughhair sampling is considered not to be invasive, it can be so forthose patients who have little hair or are indeed bald in whichcase a nonhead hair sample might be required.

SalivaSaliva was initially investigated as an alternative biolog-

ical uid for TDM of AEDs during the 1970s; the most studieddrugs are phenytoin, phenobarbital, and carbamazepine.4347

Saliva is once again emerging as a biological uid that is valu-able for AED TDM and has started to be more widely usedagain because it is associated with numerous advantages overblood/serum (Table 3). Of particular advantage is that the con-centration in saliva generally reects the free nonproteinbound pharmacologically active component in serum; salivais easier to collect than blood and patients prefer saliva sam-pling over blood sampling. Furthermore, the standard analyti-cal methods can invariably be easily adapted to accept salivaspecimens.

INDIVIDUAL AEDSIn the following section, the AEDs will be reviewed in

alphabetical order with regards to their clinical indications,key pharmacokinetic characteristics in relation to TDM ingeneral and salivary TDM in particular (Table 4), along withthe available evidence regarding the usefulness of saliva asa matrix for undertaking AED TDM.

Carbamazepine

Clinical IndicationsCarbamazepine is a rst-line drug for the treatment of

partial and secondarily generalized tonicclonic seizures andprimary generalized tonicclonic seizures. It is also the drugof choice in the management of trigeminal neuralgia and inaddition is used in the treatment of bipolar disorder that isunresponsive to lithium. Carbamazepine is available in a vari-ety of formulations including, tablets, chewable tablets, liquidoral suspension, suppositories, and extended release tabletsand capsules.

Pharmacokinetic CharacteristicsAbsorption of carbamazepine after oral ingestion is

erratic and variable with a bioavailability of 75%85% andTmax values that are formulation dependent.

81 The drug isa powerful inducer of hepatic enzymes, and after initiationof treatment, the pharmacokinetic parameters (half-life andclearance) change considerably due to autoinduction, whichgenerally is complete in about 3 weeks.82 Protein binding is75%. Carbamazepine is extensively metabolized in the liver,primarily by CYP3A4 with some contribution by CYP2C8, tocarbamazepineepoxide, which is pharmacologically active,equipotent to the parent drug and accumulates in serum to a



variable extent. Carbamazepineepoxide is subsequently metab-olized, by epoxide hydrolase, to a pharmacologically inactive10,11-diol, which is eliminated partly unchanged and partly asa glucuronide conjugate.83 Protein binding of carbamazepineepoxide is 50%60%. The serum elimination half-life (t1/2)of carbamazepine in adults is 820 hours, whereas that ofcarbamazepineepoxide is approximately 34 hours. Carbama-zepine is subject to many drug-drug pharmacokinetic interac-tions because it is an inducer of hepatic metabolism and also itsown metabolism is readily inhibited or inducedconsequentlythere are large differences between individuals in the dose toserum concentration relationship. In addition, although there isa broad range of serum concentrations associated with an opti-mum effect, there is considerable interpatient variability in theconcentration of carbamazepine that is associated with an

optimal therapeutic response (which may in part be due to thevariation in carbamazepineepoxide concentration). The fact thatthe dose to serum concentration relationship of carbamazepineand carbamazepineepoxide is nonlinear, and thus unpredict-able, is the main reason why their monitoring is useful. Thecurrent reference range for carbamazepine in serum is 412mg/L (1751 mmole/L), whereas carbamazepineepoxide con-centrations are generally 5%15% of the parent drug with a ref-erence range of up to 2.3 mg/L (up to 9 mmole/L).1

Saliva TDM for CarbamazepineThere have been many studies investigating the distribu-

tion of carbamazepine and carbamazepine epoxide in saliva andseveral have evaluated the relationship between salivary con-centration and the free, pharmacologically active, nonproteinbound concentration in serum of both adults and children withepilepsy.4854,56,73,8495

The salivary concentration of carbamazepine is similarto the free, nonprotein bound concentration in serum withmean saliva/serum total carbamazepine concentration ratiosranging 0.260.44, whereas the mean saliva/serum-free car-bamazepine concentration ratios ranged 1.391.44. Indeed,concentrations of carbamazepine in saliva are signicantlycorrelated with both serum total (r2 = 0.840.99) andserum-free carbamazepine concentrations (r2 = 0.910.99).Carbamazepineepoxide distributes into saliva such thatthe salivary concentration is similar to the free nonproteinbound concentration in serum with mean saliva/serum totalcarbamazepineepoxide concentration ratios of 0.310.55.Saliva carbamazepineepoxide concentrations are signi-cantly correlated with both serum total (r2 = 0.760.88) andserum-free (r2 = 0.750.98) carbamazepineepoxide concen-trations. Thus, saliva may be used as an alternative matrix forTDM of both carbamazepine and carbamazepineepoxide.

Clobazam

Clinical IndicationsClobazam is a 1,5-benzodiazepine drug with marked

anticonvulsant properties, which is less sedating than otherbenzodiazepines. It is licensed for use as adjunctive therapyof partial seizures or generalized seizures in patients above 3years of age and also for the management of nonconvulsivestatus epilepticus. It is also prescribed as an anxiolytic. Clobazamis available in tablet and capsule formulations.

Pharmacokinetic CharacteristicsClobazam is rapidly absorbed after oral ingestion with a

Tmax of 13 hours and bioavailability of .95%. Its pharma-cokinetics is linear and protein binding is 85%. Clobazam isextensively metabolized in the liver, primarily by CYP2C19and CYP3A4, to N-desmethyl clobazam, which is pharmaco-logically active, accumulates in serum to much higher con-centrations than the parent drug, and is responsible for muchof the clinical effect.96 N-desmethyl clobazam is subsequentlymetabolized by CYP2C19 and cleared from serum at a signif-icantly slower rate than the parent drug, with the half-life ofclobazam in adults being 1030 hours, whereas the half-lifeof N-desmethyl clobazam is 3646 hours. The protein bindingof N-desmethyl clobazam has not been reported. Clobazam is

TABLE 3. The Advantages and Disadvantages of Using Salivafor AED Therapeutic Drug Monitoring

Advantages Comments

Reects free nonprotein boundconcentration in blood

This is the ideal concentrationmeasurement in blood as it is thatcomponent (pharmacologicallyrelevant) that is accessible tothe brain where AEDs havetheir effect.

Collection is simple andnoninvasive

Avoids complications of infectionand thrombosis, which can beassociated with blood sampling.Useful for patients with needlephobias or difcult veins.

Does not require the expertiseof drawing blood

Sampling can be undertaken byspouse, partner, parent, or carer.

Cheaper than drawing blood No need for a phlebotomist, a nurse,or a doctor to bleed the patient.

Especially useful in patients withdisabilities, the elderly, andin children

Preferred by patients, parents,and carers.

Less stress, fear, and discomfort Patients are more amenable toproviding multiple samples.

Can be readily undertaken in thehome environment

Samples can be collected at theideal time (trough and predose)and readily dispatched to thehospital laboratory in advanceof the patients clinic visit.

Disadvantages Comments

Spurious results due tocontamination (drug residuesin the mouth or leakage ofdrug-rich exudate, eg, patientswith gingivitis)

This can be avoided by sampling justbefore AED ingestion (at trough),after the mouth is rinsed or aftera few hours have elapsed sincedrug ingestion.

Saliva volume insufcient This can be overcome in thelaboratory by adding distilledwater.

Difculty in pipetting due toviscosity of saliva

Sample may need to be rejected andpatient resampled.

AED concentration is low Analytical methods need to bespecically developed so as to beable to measure the anticipatedlow concentrations.

Sampling may be unacceptable Some patients may refuse sampling,although in the authorsexperience this has yet to happen.



subject to many drugdrug pharmacokinetic interactions becauseits metabolism can be readily induced or inhibitedconsequentlythere are large differences between individuals in the dose toserum concentration relationship. The current reference rangefor clobazam in serum is 0.030.3 mg/L (0.11.0 mmole/L),whereas that of N-desmethyl clobazam is 0.33.0 mg/L (1.010.5 mmole/L).1

Saliva TDM for ClobazamThere have been 2 studies investigating the distribution

of clobazam and N-desmethyl clobazam into saliva and therelationship between salivary clobazam/N-desmethyl cloba-zam concentrations and serum total clobazam/N-desmethylclobazam concentrations in children with epilepsy.54,55 Theseshowed that clobazam salivary concentration is similar to the

total concentration in serum and that concentrations are sig-nicantly correlated (r2 = 0.90). Also, N-desmethyl clobazamdistributes into saliva such that the salivary concentration issimilar to the total concentration in serum and that concen-trations are signicantly correlated (r2 = 0.93). The excellentcorrelation between total serum clobazam/N-desmethyl cloba-zam and salivary concentrations (r = 0.9 and 0.93, respec-tively) indicates that saliva can be used as an alternativematrix for clobazam and N-desmethyl clobazam TDM.

Clonazepam

Clinical IndicationsClonazepam is licensed for the treatment of a variety of

seizure types including absence, akinetic, atonic, and myo-clonic seizures. Also, it is licensed for use in patients with

TABLE 4. Pharmacokinetic Parameters and Serum Reference Ranges for the Various AEDs Prescribed as Monotherapy to Adults

AEDTime to Steady

State (d)Serum ProteinBinding (%)

Half-Life(h)

Reference Range* PharmacologicallyActive Metabolites

That Need Monitoring

Saliva MonitoringValidated

(Key References)Mg/L mole/L

Carbamazepine 24 75 820 412 1751 Carbamazepine-epoxide 4853

Clobazam 710 85 1030 0.030.3 0.11.0 N-Desmethyl-clobazam 54,55

0.33.0 1.010.5

Clonazepam 310 85 1756 0.020.07 0.060.22 Not validated

Eslicarbazepineacetate#

34 30 1320 335** 12139** Eslicarbazepine Not validated

Ethosuximide 812 0 4060 40100 283708 21,49,50,5658

Felbamate 35 25 1622 3060 126252 Not validated

Gabapentin 12 0 59 220 12117 59,60

Lacosamide 3 90 13 61

Lamotrigine 38 55 1535 2.515 1059 6264

Levetiracetam 12 0 68 1246 70270 6567

Oxcarbazepine 23 40 815 335 12139 10-Hydroxycarbazepine 6872

Phenobarbital 1530 55 70140 1040 43172 26,50,54,7375

Phenytoin 621 90 30100 1020 4079 1,50,54,56,73,76

Pregabalin 12 0 57 NE NE Not validated

Primidone 25 10 722 510 2346 Phenobarbital 49,50,56,73

Retigabine 12 80 810 NE NE Not validated

Runamide 12 35 610 3040 126168 77

Stiripentol 13 99 4.513 422 1794 Not validated

Tiagabine 12 96 59 0.020.2 0.050.53 Not validated

Topiramate 47 15 2030 520 1559 78

Valproic acid 24 90 1216 50100 346693 Salivary and serumvalproic acidconcentrationsdo not correlate.

Vigabatrin 12 0 58 0.836 6279 79

Zonisamide 912 40 5070 1040 47188 80

*For clarity, values can be rounded up or down by the laboratory.Refers to patients on chronic therapy after autoinduction is completedvalues are much greater after a single dose.There are clinical settings where monitoring of carbamazepine-epoxide, in addition to carbamazepine, is warranted, particularly when a comedication occurs with an inhibitor ofcarbamazepineepoxide metabolism.Includes time to steady state for active metabolite N-desmethyl clobazam.Refers to values for active metabolite N-desmethyl-clobazam.#All values refer to the active metabolite eslicarbazepine.**The reference range is that quoted for the active metabolite of oxcarbazepine, namely, 10-hydroxycarbazepine because the 2 molecules are identical.All values refer to the active metabolite 10-hydroxycarbazepine.Elimination is saturable so that half-live increases with increasing plasma concentration.Not established.During treatment with primidone both primidone and the pharmacologically active metabolite phenobarbital should be monitored.



LennoxGastaut syndrome and in the management of statusepilepticus. Clonazepam is available in a variety of formula-tions including tablets, a disintegrating wafer and a liquidformulation for intravenous administration.

Pharmacokinetic CharacteristicsClonazepam is rapidly absorbed after oral ingestion with a

Tmax of 14 hours and bioavailability of.80%. It exhibits linearpharmacokinetics and protein binding is 85%. Clonazepam isextensively metabolized in the liver, primarily by CYP3A4, to7-aminoclonazepam, which in turn is metabolized by acetyla-tion, via N-acetyl-transferase, to form 7-acetamidoclonazepam.The serum elimination half-life of clonazepam in adults is 1756hours so that interindividual clearance is extremely variable.9799

Clonazepam is subject to some drugdrug pharmacokineticinteractions consequent to the fact that its metabolism is readilyinduced or inhibitedtherefore, large differences exist betweenindividuals in the dose to serum concentration relationship.The current reference range for clonazepam in serum is 0.020.07 mg/L (0.060.22 mmole/L).1

Saliva TDM for ClonazepamIt is not known whether clonazepam is secreted into

saliva and if it is whether the concentrations reect those inserum. Analysis of saliva samples spiked with clonazepam andstored overnight at room temperature resulted in concentrationsthat were 76% lower compared with spiked saliva samples thatwere analyzed immediately.100 These data suggest that clona-zepam is unstable in saliva. Interestingly, clonazepam spikedinto water was stable.100

Eslicarbazepine Acetate

Clinical IndicationsEslicarbazepine acetate is licensed for the adjunctive

treatment of partial onset seizures with or without secondarygeneralization in patients with epilepsy aged 16 years andolder. The drug is available as formulations of tablet and asuspension.

Pharmacokinetic CharacteristicsEslicarbazepine acetate is a prodrug, and after oral absorp-

tion, the acetate group is rapidly and extensively metabolized byhydrolytic rst pass metabolism with esterases to eslicarbaze-pine (S-licarbazepine), the S-enantiomer of the pharmacologi-cally active 10-hydroxycarbazepine metabolite of oxcarbazepine(also known as the monohydroxy derivative). After oral inges-tion, eslicarbazepine acetate is rapidly absorbed with a Tmax foreslicarbazepine of 23 hours and bioavailability of .90%.101

Eslicarbazepine pharmacokinetics is linear and protein bindingis 30%. In addition to eslicarbazepine, small amounts of 2 otherpharmacologically active metabolites are formed from eslicar-bazepine acetate (R-licarbazepine and oxcarbazepine), but theserepresent only approximately 6% of metabolites. Eslicarbaze-pine and its glucuronides together with minor quantities ofR-licarbazepine, oxcarbazepine, eslicarbazepine acetate, andtheir respective glucuronides are excreted in urine. Eslicarba-zepine glucuronic acid conjugation is primarily catalyzed byUGT1A4, UGT1A9, UGT2B4, UGT2B7, and UGT2B17.102

The serum half-life of eslicarbazepine in adults is 1320hours, and it is subject to very few drugdrug pharmacoki-netic interactions.103 The current reference range for eslicar-bazepine in serum is 335 mg/L (12139 mmole/L), which isbased on that for racemic 10-hydroxycarbazepine derived fromoxcarbazepine.1

Saliva TDM for Eslicarbazepine AcetateIt is not known whether eslicarbazepine acetate is secreted

into saliva or whether salivary concentrations are similar orreect those in serum. However, because the pharmacologi-cally active metabolite, eslicarbazepine, is the same moleculeas the pharmacologically active metabolite of oxcarbazepine,10-hydroxycarbazepine, it can be expected that its transfer intosaliva will be similar to that described for 10-hydroxycarbaze-pine in Oxcarbazepine.

Ethosuximide

Clinical IndicationsEthosuximide is licensed for monotherapy treatment of

absence seizures in patients of all ages and is available asformulations of capsule and a syrup.

Pharmacokinetic CharacteristicsEthosuximide is rapidly absorbed after oral ingestion

with a Tmax of 14 hours and its bioavailability is .90%. Itspharmacokinetics is linear and it is not protein bound.104,105

Ethosuximide is extensively metabolized in the liver, primar-ily by CYP3A and to a lesser extent by CYP2E and CYP2B/C, to form isomers of 2-(1-hydroxymethyl)-2-methylsuccina-mide of which 40% are excreted as glucuronide conjugates.The serum half-life of ethosuximide in adults is 4060 hourswith large interindividual differences in serum clearance. Fur-thermore, ethosuximide is subject to a number of drugdrugpharmacokinetic interactions because its metabolism can beboth induced and inhibited with consequent large differencesbetween individuals in the dose to serum concentration rela-tionship. The current reference range for ethosuximide inserum is 40100 mg/L (283708 mmole/L).1

Saliva TDM for EthosuximideThere have been several studies investigating the distri-

bution of ethosuximide into saliva and the relationship betweensaliva ethosuximide concentrations and total serum ethosuximideconcentrations in patients with epilepsy.21,49,50,5658 Ethosuximideis not protein bound and the salivary concentration is similar tothe total serum concentration with mean saliva/serum total etho-suximide concentration ratios ranging 0.951.04. Indeed, salivaand serum total ethosuximide concentrations are signicantlycorrelated (r2 = 0.99), and thus, saliva can be used as an alter-native matrix for ethosuximide TDM.

Felbamate

Clinical IndicationsBecause felbamate is associated with an increased risk

of aplastic anemia and hepatotoxicity, its use is restricted suchthat it is approved for use only in patients who respond inade-quately to alternative treatments and particularly in patients with



partial seizures or LennoxGastaut syndrome. Felbamate isavailable as formulations of tablet and a suspension.

Pharmacokinetic CharacteristicsFelbamate is rapidly absorbed after oral ingestion with

a Tmax of 26 hours and a bioavailability of .90%. It haslinear pharmacokinetics and protein binding is 25%. About50% of an administered dose is metabolized in the liver,primarily by CYP3A4 and CYP2E1, to form 2 hydroxylatedmetabolites (p-hydroxy and 2-hydroxy felbamate). In additionone of the carbamate groups is hydrolyzed to an alcohol thatis further biotransformed to an acid; also a number of as yetunidentied polar metabolites are produced, some of whichare glucuronides.106,107 The development of hepatotoxicityand aplastic anemia in a few patients treated with felbamateis due to the formation of a reactive atropaldehyde metabolite,which can accumulate in some patients and cause toxicity.106

The serum half-life of felbamate in adults is 1622 hours.Felbamate is subject to drugdrug pharmacokinetic interac-tions consequent to the fact that its metabolism is both readilyinduced and inhibited; also felbamate itself acts as an inhibitorof hepatic metabolism.108110 Consequently, there are large dif-ferences between individuals in the dose to serum concentra-tion relationship. The current reference range for felbamate inserum is 3060 mg/L (126252 mmole/L).1

Saliva TDM for FelbamateIt is not known whether felbamate is secreted into saliva

and if it is whether concentrations are similar to or reect thosein serum.

Gabapentin

Clinical IndicationsGabapentin is licensed for the monotherapy treatment

of partial seizures with or without secondary generalization inadults and children aged 12 years and above, and as an adjunc-tive treatment in adults and children aged 6 years and above.The drug is also licensed for the treatment of peripheral neu-ropathic pain. Gabapentin is available as formulations of tabletsand capsules.

Pharmacokinetic CharacteristicsGabapentin is rapidly absorbed after oral ingestion with

a Tmax of 23 hours. Bioavailability is 60% and is dose depen-dent with bioavailability decreasing at higher doses. Its pharma-cokinetics is nonlinear consequent to its saturable absorptionfrom the proximal small bowel primarily by the L-amino acidtransport system.111 Gabapentin is not protein bound and notmetabolized, being cleared entirely by renal excretion witha serum elimination half-life in adults of 59 hours.112 Neverthe-less, although gabapentin is not subject to drugdrug pharma-cokinetic interactions,103 the disposition can be extremelyvariable because of wide interindividual differences is absorp-tion.113 The fact that gabapentin is associated with nonlinearpharmacokinetics is a major reason why gabapentin monitoringis particularly valuable for patient management and thecurrent reference range for gabapentin in serum is 220 mg/L(12117 mmole/L).1

Saliva TDM for GabapentinThere have been 3 studies investigating the distribution

of gabapentin in saliva1 in healthy volunteers59 and 2 inpatients with epilepsy.60,114 Gabapentin distributes into saliva;however, mean concentrations are 2.410% of those observedin serum.59,60 Nevertheless, there is signicant correlation (r2 .0.7) between saliva and serum total gabapentin concentra-tions114; furthermore, salivary gabapentin concentration and doseare signicantly correlated (r2 = 0.770.95).59,60 Thus, salivamay be a useful alternative matrix for gabapentin TDM.

Lacosamide

Clinical IndicationsLacosamide is licensed for the adjunctive treatment of

partial onset seizures with or without secondary generaliza-tion in patients with epilepsy aged 16 years and older.Lacosamide is available as formulations of tablets, a solution,and a syrup.

Pharmacokinetic CharacteristicsLacosamide is rapidly absorbed after oral ingestion with

a Tmax of 12 hours, and its bioavailability is 100%.115 Its

pharmacokinetics is linear and protein binding is controver-sial with Greenaway et al61 reporting 90% and the Summaryof Product Characteristic116 stating that serum lacosamideprotein binding is,15%. About 60% of a dose of lacosamideis hepatically metabolized, by demethylation via CYP2C19,to form O-desmethyl lacosamide.102 The serum eliminationhalf-life of lacosamide in adults is 13 hours115 and to dateno drugdrug pharmacokinetic interactions have been identi-ed.103 The current reference range for lacosamide in serum is1020 mg/L (4080 mmole/L).

Saliva TDM for LacosamideThere has been one study investigating the distribution

of lacosamide into saliva and the relationship between salivalacosamide concentration and both serum total and free concen-trations in adults with epilepsy.61 Lacosamide distributes intosaliva such that the salivary concentration is similar to the nonprotein bound concentration in serum; mean saliva/serum-freelacosamide concentration ratios ranged 0.770.96. Furthermore,saliva lacosamide concentrations are signicantly correlatedwith both serum total lacosamide (r2 = 0.84) and serum-freelacosamide (r2 = 0.83) concentrations. Thus, saliva should bea useful alternative matrix for lacosamide TDM.

Lamotrigine

Clinical IndicationsLamotrigine is licensed for the monotherapy treatment

of partial seizures and primary and secondarily generalizedtonicclonic seizures in adults and children over 12 years ofage; as adjunctive treatment in adults and children over 2years of age; as adjunctive treatment of seizures associatedwith the LennoxGastaut syndrome in adults and childrenover 2 years of age. Lamotrigine is also licensed for thetreatment of bipolar I disorder. It is available as formulationsof tablets and dispersible chewable tablets.



Pharmacokinetic CharacteristicsLamotrigine is rapidly absorbed after oral ingestion

with a Tmax of 13 hours and its bioavailability is .95%.117

Its pharmacokinetics is linear and protein binding is 55%.Lamotrigine is extensively metabolized in the liver, primarilyby glucuronidation via UGT1A4, to form N-2 and N-5 glu-curonides.117 The serum elimination half-life of lamotrigine inadults is 1535 hours so that interindividual clearance isextremely variable118; furthermore, it is subject to manydrugdrug pharmacokinetic interactions consequent to thefact that its metabolism can be both induced and inhibited;consequently, there are large differences between individualsin the dose to serum concentration relationship.103 The currentreference range for lamotrigine in serum is 2.515 mg/L(1059 mmole/L).1

Saliva TDM for LamotrigineLamotrigine is reported to be about 55% bound to serum

proteins in patients receiving 150300 mg/d in conjunctionwith other medication and the saliva/serum lamotrigine ratiois reported to be 0.46 in healthy subjects receiving a singledose and 0.56 in patients receiving adjunctive therapy.119,120

The excellent correlation between serum and saliva concentra-tions of lamotrigine (r = 0.95) in these early studies suggestedthat saliva could potentially be used to monitor the systemicconcentrations of lamotrigine.

A subsequent study examined the interindividual corre-lation between lamotrigine concentrations in saliva and serumtogether with the relationship between saliva concentrationand the nonprotein bound lamotrigine concentration inserum.62 The authors compared both stimulated and unstimu-lated saliva from the same patients and demonstrated a goodcorrelation between lamotrigine serum concentration in bothcollection modes (r2 = 0.85, unstimulated and r2 = 0.94,stimulated). Furthermore, the study demonstrated a good cor-relation between total lamotrigine concentration in serum andthe free concentration as determined by ultraltration (r2 = 0.95)and equilibrium dialysis (r2 = 0.93). Lamotrigine concentrationin stimulated saliva was also signicantly correlated with thefree concentration and calculation of lamotrigine protein bind-ing using the 3 alternative procedures gave the following results(mean 6 SD); 51.8% 6 13.03% (stimulated), 68.05% 6 7.59%(ultraltration), and 58.72%6 7.68% (equilibrium dialysis). Thedifferences in calculated binding between the 3 methods weresignicant.62

Ryan et al63 studied the relationship between serum andsalivary concentrations of lamotrigine in both pediatric andadult epilepsy populations and reported a good correlationbetween the two (r2 = 0.810.84) and with the saliva/serumlamotrigine concentration ratios ranging 0.401.19 (mean 6SD = 0.64 6 0.18). The authors concluded that althougha good correlation existed for the population at large betweensalivary and serum concentrations for lamotrigine, there iswide interpatient variability in the saliva/serum ratio. The datasuggest that salivary monitoring may play a role in the mon-itoring of lamotrigine for adult and pediatric patients.

In another study, lamotrigine concentrations weremeasured in both stimulated and unstimulated saliva along-side matching serum samples from 7 adult volunteers over

a 32-hour period after a single 50-mg dose of the drug, also insamples from 20 children and adolescents during the courseof routine AED therapy.64 In specimens collected $2 hoursafter ingestion, there was a close correlation in each individ-ual between the concentrations in stimulated and unstimulatedsaliva, which were similar. The saliva/serum lamotrigine con-centration ratio gave a mean value of 0.49 at a serum lamo-trigine concentration of 10 mg/L, and the authors concludedthat with appropriate precautions regarding the timing of sam-ple collection saliva measurements could provide a reasonablealternative to serum for TDM.64

A study in 14 healthy volunteers comparing saliva andserum lamotrigine concentrations over 96 hours after inges-tion of a single oral dose of lamotrigine reported signicantcorrelation between saliva and serum (r2 = 0.677).121 Further-more, the mean saliva/serum lamotrigine concentration ratiowas 0.425 6 0.153, and the calculated protein binding fromthe concentration in saliva was 57.5 6 15.1% (mean 6 SD);thus, saliva concentrations reect the free concentrations inserum. More recently, Mallayasamy et al122 reported a corre-lation between salivary and serum lamotrigine concentrationsof 0.683.

In summary, for lamotrigine, there are many factors thatmake TDM clinically useful, and several studies have foundgood correlation between salivary concentrations and both totaland free nonprotein bound serum concentrations; thus, lamo-trigine TDM in saliva is a viable alternative to that of serum.

Levetiracetam

Clinical IndicationsLevetiracetam is licensed for the monotherapy treatment

of partial seizures with or without secondary generalization inpatients aged 16 years and older and as adjunctive treatment inadults and children from 4 years of age. The drug is alsolicensed for the adjunctive treatment of primary generalizedtonicclonic seizures associated with idiopathic generalizedepilepsy and myoclonic seizures associated with juvenile myo-clonic epilepsy in adults and adolescents from 12 years of age.Levetiracetam is available in a variety of formulations includ-ing tablets, an oral solution, a solution for intravenous injec-tion, and an extended release tablet formulation.

Pharmacokinetic CharacteristicsLevetiracetam is rapidly absorbed after oral ingestion

with a Tmax of 12 hours and a bioavailability of .95%. Itspharmacokinetics is linear, and it is not protein bound.123

Approximately 30% of a dose of levetiracetam undergoesmetabolism by a cytosolic amidase enzyme to produce a car-boxylic acid metabolite (2-pyrrolidone-N-butyric acid), whichis excreted unchanged via the kidneys.124,125 Levetiracetammetabolism to the carboxylic acid is independent of the hepaticCYP system and occurs by means of a type-B esterase locatedin whole blood.126 However, the drug also undergoes a smallamount of hepatic metabolism to form 2 ring-hydroxylatedmetabolites. The serum elimination half-life of levetiracetamin adults is 68 hours, and it is subject to minimal drugdrugpharmacokinetic interactions.103 The current reference rangefor levetiracetam in serum is 1246 mg/L (70270 mmole/L).1



Saliva TDM for LevetiracetamThere have been several studies investigating the distri-

bution of levetiracetam into saliva and the relationship betweensaliva levetiracetam and serum concentrations in adults andchildren with epilepsy and also in healthy volunteers.6567

Levetiracetam distributes into saliva. However, there issome controversy regarding whether or not saliva and serumlevetiracetam concentrations are the same. Grim et al65 reportthat salivary concentrations are approximately 40% of thatobserved in serum, with mean saliva/serum levetiracetam con-centration ratios ranging 0.360.41. In contrast, Lins et al66 andMecarelli et al67 report that the concentration of levetiracetamin saliva is similar to serum with the mean saliva/serum leve-tiracetam concentration ratio being 1.0 and 1.1, respectively.Interestingly, a ratio of 1.55 was observed after ingestion of anoral solution in healthy volunteers.66 Nevertheless, saliva andserum levetiracetam concentrations are signicantly correlated(r2 = 0.860.91) although lemon juice stimulation reduces thecorrelation from r2 = 0.91 to r2 = 0.87. Overall, the data sug-gest that saliva may be used as an alternative matrix for leve-tiracetam TDM.

Oxcarbazepine

Clinical IndicationsOxcarbazepine is licensed for the monotherapy or adjunc-

tive treatment of partial seizures with or without secondarygeneralization in patients aged 6 years or more. It is available asformulations of tablets and an oral suspension.

Pharmacokinetic CharacteristicsOxcarbazepine is a prodrug and is rapidly metabolized,

by cytosolic arylketone reductase, to a pharmacologicallyactive metabolite 10-hydroxycarbazepine (also known aslicarbazepine or monohydroxy metabolite). This metaboliteaccumulates in serum and is responsible for most of the drugeffects. The conversion of oxcarbazepine to 10-hydroxycarba-zepine is stereoselective and concentrations of the S-enantio-mer are somewhat higher than those of the R-enantiomer.127,128

After oral ingestion, oxcarbazepine is rapidly absorbed witha Tmax of 36 hours and a bioavailability of 100%. Oxcarba-zepine pharmacokinetics is linear and protein binding is 60%,whereas 10-hydroxycarbazepine protein binding is 40%.129 10-Hydroxycarbazepine is subsequently metabolized by conjuga-tion with glucuronic acid, and the conjugates together withsome 10-hydroxycarbazepine are excreted in urine. The serumelimination half-life of 10-hydroxycarbazepine in adults is 815 hours, and it is subject to many drugdrug pharmacokineticinteractions because its metabolism is both readily inhibitedand induced. Oxcarbazepine is also itself a weak inducer ofhepatic metabolism130,131; consequently, there are large differ-ences between individuals in the dose to serum concentrationrelationship. The current reference range for 10-hydroxycarba-zepine in serum is 335 mg/L (12139 mmole/L).1

Saliva TDM for 10-HydroxycarbazepineThere have been several studies investigating the distri-

bution of 10-hydroxycarbazepine into saliva and the rela-tionship between salivary concentrations and serum total/free

10-hydroxycarbazepine concentrations in healthy volunteersand in adults and children with epilepsy.6872 10-Hydroxycar-bazepine distributes into saliva; however, the extent dependson whether resting saliva or stimulated saliva is collected.Thiesohn and Heimann68 in a study of 3 healthy volunteersobserved that in resting saliva the saliva/serum ratio ranged0.31.7 (median 1.0). Kristensen et al69 in a study of 7 healthyvolunteers reported that in stimulated saliva (parafn wax) themean saliva/serum ratio was 0.53. The study of Klitgaard andKristensen70 involving 17 patients with epilepsy reported thatin resting saliva the mean saliva/serum ratio was 1.01, whereasin stimulated saliva (parafn wax), the value was 0.41. Cardotet al71 in a study of 10 patients with epilepsy reported that instimulated saliva (citric acid) the mean saliva/serum ratio was0.19. These data indicate that with greater saliva stimulationthe saliva/serum ratio decreases such that saliva 10-hydroxy-carbazepine concentrations approached the free 10-hydroxycar-bazepine concentrations in serum. This may be due to therelatively low lipid solubility of 10-hydroxycarbazepine; a char-acteristic that is not associated with carbamazepine. Unfortu-nately, increasing salivary ow has an extremely variable effecton saliva 10-hydroxycarbazepine concentration, which resultsin the saliva and serum 10-hydroxycarbazepine concentrationsnot being signicantly correlated. However, when unstimulatedsaliva is collected, there is a signicant correlation betweensaliva and serum total 10-hydroxycarbazepine concentrations(r2 = 0.910.98).6870 Indeed, a recent study by Miles et al72 of28 children and adult patients with epilepsy reported on un-stimulated saliva whereby the mean saliva/serum ratio was0.96 and saliva 10-hydroxycarbazepine concentrations weresignicantly correlated with serum total 10-hydroxycarbazepineconcentrations (r2 = 0.94). Thus unstimulated saliva may be auseful alternative matrix for 10-hydroxycarbazepine TDM.

Phenobarbital

Clinical IndicationsPhenobarbital is licensed for the monotherapy or adjunc-

tive treatment of all forms of epilepsy, except absence seizures,in patients of any age. Phenobarbital is available in a varietyof formulations including, tablets, a solution for intravenousinjection, and an elixir formulation.

Pharmacokinetic CharacteristicsPhenobarbital is rapidly absorbed after oral ingestion

with a Tmax of 24 hours and a bioavailability of .90.132,133

Its pharmacokinetics is linear, and protein binding is 55%.Phenobarbital is extensively metabolized in the liver, primarilyby CYP2C9 and to a lesser extent by CYP2C19 and CYP2E1,to form 2 major metabolites, p-hydroxyphenobarbital and a9-D-glucopyranosyl phenobarbital isomer. The serum half-lifeof phenobarbital in adults is 70140 hours so that interindivid-ual clearance is extremely variable132,133 Phenobarbital is subjectto many drugdrug pharmacokinetic interactions consequent tothe fact that it is a potent inducer of hepatic metabolism and alsoits own metabolism can be induced or inhibited. Consequently,there are large differences between individuals in the dose toserum concentration relationship. The current reference rangefor phenobarbital in serum is 1040 mg/L (43172 mmole/L).1



Saliva TDM for PhenobarbitalUsing saliva as an alternative matrix for TDM of pheno-

barbital is somewhat controversial because there is no clearconsensus regarding whether the concentration in saliva directlyreects the concentrations of the drug in serum. A number ofstudies in both adults and children with epilepsy have demon-strated that phenobarbital distributes into saliva with saliva/totalserum phenobarbital concentration ratios ranging 0.20.52,whereas the mean saliva/serum-free phenobarbital concentrationratios ranged 0.630.68. Indeed, saliva phenobarbital concentra-tions and both serum total phenobarbital (r2 = 0.650.98) andserum-free phenobarbital (r2 = 0.640.99) concentrations aresignicantly correlated.26,49,50,53,54,56,57,7375,134137

Two studies have demonstrated that the distributionof phenobarbital into saliva depends on salivary pH56,136;however, other studies have not found an effect of pH.75,137

The study by McAuliffe et al56 determined phenobarbital insaliva and serum obtained simultaneously from 115 pa-tients, and a method to correct for the effect of salivarypH on drug concentration of saliva was developed. Salivaryphenobarbital concentration was found to be equivalentto the free phenobarbital concentration in serum and tocorrelate signicantly with the total serum concentration.Expressed as percentage of total serum drug, the salivary(S) and serum-free (P) concentrations were: phenobarbital,S 43.1 6 5.2%, P = 40.8 6 7.9% (r = 0.91). On balance, itseems that salivary concentrations of phenobarbital corre-late with the simultaneous serum water concentrations, aftercorrecting for the effects of pH differences between salivaand serum.138

Despite the contradictory results, when equations incor-porate the relative ratio of phenobarbital pKa to the salivarypH, there is an excellent correlation between salivary and freenonprotein bound serum phenobarbital concentrations. Thus,saliva can be regarded as a useful alternative matrix for phe-nobarbital TDM.

Phenytoin

Clinical IndicationsPhenytoin is licensed for both monotherapy and adjunc-

tive therapy of clonicclonic seizures and focal seizures inpatients of any age. It is also approved for the treatment ofseizures occurring during or after neurosurgery and/or severehead injury. In addition, it is licensed for intravenous admin-istration in the management of established status epilepticusand for monotherapy use in the treatment of trigeminal neu-ralgia. Phenytoin is available in a variety of formulationsincluding, capsules, chewable tablets, an oral suspensionand a parenteral solution formulation. More recently, fosphe-nytoin has been licensed, which is a water-soluble phenytoinphosphate prodrug that is rapidly dephosphorylated on admin-istration. It is formulated for intravenous administration tocontrol of status epilepticus and the prevention/treatment ofseizures occurring in connection with neurosurgery and/orhead trauma. It can also be substituted for oral phenytoinif oral administration is not possible and/or contraindicated.Fosphenytoin is formulated as a solution for infusion orinjection.

Pharmacokinetic CharacteristicsThe rate of absorption of phenytoin after oral ingestion

is variable and formulation dependent with a Tmax of 112hours.139 Bioavailability is similarly formulation dependentbut is .80%. Its pharmacokinetics is nonlinear due to satu-rable metabolism, which results in MichaelisMenten kineticswithin the range of serum concentrations that are generallyassociated with its benecial therapeutic effects.140 The pro-tein binding of phenytoin is 90%. Phenytoin is extensivelymetabolized in the liver, primarily by CYP2C9 andCYP2C19, to form 2 major metabolites 5-(p-hydroxyphen-yl)-5-phenylhydantoin (which undergoes partial conversionto glucuronides before renal excretion) and a dihydrodiolderivative.141 The serum elimination half-life of phenytoinin adults is 30100 hours so that interindividual clearance isextremely variable. Phenytoin is subject to many drugdrugpharmacokinetic interactions consequent to the fact that it isa potent inducer of hepatic metabolism; also its metabolismcan be both induced and inhibitedconsequently, there arelarge differences between individuals in the dose to serumconcentration relationship. The fact that phenytoin is associ-ated with nonlinear pharmacokinetics is one of the majorreasons why monitoring the drug is so useful. The currentreference range for phenytoin in serum is 1020 mg/L (4079 mmole/L).1

Saliva TDM for PhenytoinThere have been many studies investigating the distri-

bution of phenytoin into saliva and the relationship betweensaliva phenytoin concentration and both serum total and freephenytoin concentrations in both adults and children withepilepsy.26,49,53,54,56,57,73,76,90,95,134,135,137,142146 Phenytoin dis-tributes into saliva such that the salivary concentration issimilar to the free nonprotein bound concentration in serum.Mean saliva/serum total phenytoin concentration ratiosranged 0.090.13, whereas the mean saliva/serum-free phe-nytoin concentration ratios ranged 0.991.06. Indeed, salivaphenytoin concentrations and both serum total phenytoin(r2 = 0.850.99) and serum-free phenytoin (r2 = 0.960.99)concentrations are signicantly correlated. There is a sugges-tion that the extent of phenytoin distribution in saliva dependson whether resting saliva, stimulated saliva, or reduced owsaliva is collected. For the latter 2 situations, phenytoin concen-trations in saliva are decreased and increased, respectively.147

Thus, unstimulated saliva should be used, and this provides anextremely useful alternative matrix for phenytoin TDM.

Pregabalin

Clinical IndicationsPregabalin is licensed as adjunctive treatment of partial

seizures with or without secondary generalization in adults. Itis also licensed for the treatment of peripheral and central painand for generalized anxiety disorders. Pregabalin is availableas formulations of capsules only.

Pharmacokinetic CharacteristicsPregabalin is rapidly absorbed after oral ingestion

with a Tmax of 12 hours and bioavailability of .90%.148 Its



pharmacokinetics is linear, and it is not protein bound. Prega-balin is not metabolized being cleared entirely by renal excre-tion with a serum elimination half-life of 57 hours in adults.148

Pregabalin is not subject to drugdrug pharmacokinetic inter-actions.103 The precise role for TDM of pregabalin has not yetbeen established, although there may be a requirement inpatients with renal impairment, to ascertain compliance, wheremalabsorption is suspected and in cases of suspected overdose.Very little information is available regarding therapeutic serumconcentrations of pregabalin; however, one report states that insamples collected at random times relative to dose frompatients maintained on 600 mg/d, serum pregabalin concentra-tions ranged from 0.9 to 14.2 mg/L. There are reports of sig-nicant toxicity in a case of self-poisoning with pregabalinalone149; also, a case of toxicity associated with therapeuticuse in a patient with renal failure although the peak drug con-centration (predialysis) in this case was only 13 mg/L.

Saliva TDM for PregabalinIt is not known whether pregabalin is secreted into

saliva and if it is whether concentrations are similar to orreect those in serum.

Primidone

Clinical IndicationsPrimidone is licensed for the monotherapy or poly-

therapy treatment of generalized tonicclonic seizures, psy-chomotor and focal seizures in adults and children and for themanagement of Jacksonian seizures, myoclonic jerks, andakinetic attacks. It is also licensed for the treatment of essen-tial tremor. Primidone is available as tablet formulations anda suspension.

Pharmacokinetic CharacteristicsPrimidone is rapidly absorbed after oral ingestion

with a Tmax of 24 hours and bioavailability of .90%.150

Its pharmacokinetics is linear and protein binding is 10%.Primidone is extensively metabolized in the liver to form 2major pharmacologically active metabolites, phenobarbitaland phenyl-ethyl-malonamide. The serum elimination half-life of primidone in adults is 722 hours so that interindivid-ual clearance is variable.150 Primidone is subject to manydrugdrug pharmacokinetic interactions consequent to the factthat it is a potent inducer of hepatic metabolism (via pheno-barbital). Also, its metabolism can be induced or inhibited;consequently, there are large differences between individualsin the dose to serum concentration relationship. During treat-ment with primidone, it is common practice to monitor pheno-barbital because the adverse effects from a high phenobarbitalconcentration is more likely to limit a primidone dosageincrease. The current reference range for primidone in serumis 510 mg/L (2346 mmole/L).1

Saliva TDM for PrimidoneThere have been several studies that investigate the

distribution of primidone in saliva and the relationship betweensaliva primidone and serum total primidone concentrations inboth adults and children with epilepsy.26,49,50,56,57,73,151 Themean saliva/serum total primidone concentration ratios ranged

0.071.15. Furthermore, saliva and total serum primidone con-centrations are signicantly correlated (r2 = 0.710.97). Sali-vary primidone concentrations seem to be ow dependent withresting saliva concentrations being approximately 38% lowerthan in ow-stimulated saliva.152 Thus, provided that samplecollection is standardized saliva may be used as an alternativematrix for primidone TDM.

Retigabine

Clinical IndicationsRetigabine is indicated as adjunctive treatment of partial

onset seizures with or without secondary generalization in adultsaged 18 years and above with epilepsy. Retigabine is availableas formulations of tablet only.

Pharmacokinetic CharacteristicsRetigabine is rapidly absorbed after oral ingestion with

a Tmax of 0.61.5 hours and bioavailability of about 60%.152

Its pharmacokinetics is linear and protein binding is 80%.153

Approximately 20%30% of the administered dose of retiga-bine is eliminated renally unchanged with the remaining drugbeing biotransformed to produce the N-acetyl metabolite(which has a weak pharmacological action) and N-glucuro-nide conjugates of both parent drug and N-acetylretigabineaccumulates in plasma reaching similar concentrations to theparent drug and while the N-glucuronide metabolites are phar-macologically inactive, they may contribute to the enterohe-patic circulation of retigabine and N-acetylretigabine.102 Theserum elimination half-life of retigabine in adults is 810hours, and it is subject to only a few drugdrug pharmacoki-netic interactions.103 The precise role for TDM of retigabinehas not yet been established, although there may be a require-ment in patients to ascertain compliance, where malabsorp-tion is suspected and in cases of suspected toxicity. There areno data relating plasma retigabine levels with that of seizuresuppression or adverse effects.

Saliva TDM for RetigabineIt is not known whether retigabine is secreted into

saliva and if so whether concentrations are similar to or reectthose in serum.

Rufinamide

Clinical IndicationsRunamide is licensed for the adjunctive treatment of

seizures associated with the LennoxGastaut syndrome inpatients 4 years and older. Runamide is available as formu-lations of tablet only.

Pharmacokinetic CharacteristicsRunamide is rapidly absorbed after oral ingestion with

a Tmax of 46 hours and bioavailability decreases withincreasing dose. Consequently, its pharmacokinetics is linearup to only 1600 mg/d. Protein binding is 35%.154,155 Becausefood coingestion enhances Cmax values by 50% and increasesarea under the concentration versus time curve values by33%, probably by improving the solubility of runamide,patients should be advised to take their runamide dose each



time in the same temporal relation to their meals to maintainsteady-state concentrations from one dose to the next.156 Ru-namide is extensively metabolized in the liver, primarily byan amidase hydrolysis (which is not CYP dependent) thatconverts the carboxamide function to the corresponding car-boxylic acid, which is not pharmacologically active.154,157 Theacid subsequently undergoes glucuronidation before renalexcretion. The serum elimination half-life of runamide inadults is 610 hours.158 Runamide is subject to somedrugdrug pharmacokinetic interactions consequent to thefact that its metabolism can be both induced andinhibited103consequently, there are large differences betweenindividuals in the dose to serum concentration relationship.Furthermore, because runamide is associated with nonlinearpharmacokinetics, there is an excellent rational for runamideTDM. The current information indicates that serum concentra-tions in the range of 3040 mg/L (126168 mmole/L) arerequired for seizure control in patients with LennoxGastautsyndrome; however, lower concentrations may prove to beeffective in other seizure types.154,155

Saliva TDM for RunamideRunamide distributes into saliva and preliminary

measurements in a single patient of saliva and serum samplesat steady state after 3 different runamide doses indicate thatthe mean saliva to serum concentration ratio was 0.66.77

Because the serum protein binding of runamide is 35%, thesedata indicate that salivary runamide concentrations reect thenonprotein bound drug concentrations in serum.154 However,a larger population of patients will require testing to conrmthis preliminary nding and also a direct comparison betweenthe concentration in saliva and the unbound concentration inserum will need to be performed. Nevertheless, it seems thatTDM of runamide in saliva is likely to be a useful alternativeto serum, particularly if salivary concentrations of runamidereect the free, pharmacologically active concentration in serum.

Stiripentol

Clinical IndicationsStiripentol is licensed for the adjunctive treatment of

seizures in children with severe myoclonic epilepsy in infancy(Dravet syndrome). Stiripentol is available as formulations ofcapsule- and sachet-containing granules.

Pharmacokinetic CharacteristicsStiripentol is rapidly absorbed after oral ingestion with

a Tmax of 0.52.0 hours; however, its bioavailability has yet tobe determined.159 Stiripentol pharmacokinetics is nonlineardue to saturable metabolism, which results in MichaelisMenten kinetics within the range of serum concentrations thatare generally associated with benecial therapeutic effects.160

Protein binding is 99%, and the drug is subject to extensiverst pass metabolism.159 The serum elimination half-life ofstiripentol in adults is 4.513 hours, and its metabolism iscomplex with 13 different metabolites having been identi-ed.161,162 The principal hepatic enzymes involved are CY-P1A2, CYP2C19, and CYP3A4 and interindividual clearancesof stiripentol are extremely variable. Stiripentol is subject to

many drugdrug pharmacokinetic interactions consequent tothe fact that it is a potent inhibitor of hepatic metabolism andalso its own metabolism can be induced103; therefore, largedifferences occur between individuals in the dose to serumconcentration relationship. The nonlinear pharmacokinetics ofstiripentol is the primary reason why monitoring is useful.Although the reference range for stiripentol in serum is notwell dened, concentrations of 422 mg/L (1794 mmole/L)correlate with control of absence seizures in children163 and inDravet syndrome concentrations of 812 mg/L (3451 mmole/L)are reported to be effective.164

Saliva TDM for StiripentolIt is not known whether stiripentol is secreted into

saliva and if it is whether concentrations are similar to orreect those in serum. However, if the concentration reectsthe free nonprotein bound concentration it will present ananalytical challenge because of the 99% binding of stiripentol.

Tiagabine

Clinical IndicationsTiagabine is licensed for the adjunctive treatment of

partial seizures with or without secondary generalization inadults and children aged 12 years and above. Tiagabine isavailable as formulations of tablet only.

Pharmacokinetic CharacteristicsTiagabine is rapidly absorbed after oral ingestion with

a Tmax of 0.52 hours and bioavailability is .90%. Its phar-macokinetics is linear and protein binding is 96%. Tiagabineis extensively metabolized in the liver, primarily by CYP3A4,to form two 5-oxo-tiagabine isomers together with someadditional minor metabolites.165 The serum elimination half-life of tiagabine in adults is 59 hours. The drug is subject todrugdrug pharmacokinetic interactions consequent to thefact its metabolism can be readily induced103; therefore, largedifferences occur between individuals in the dose to serumconcentration relationship.166 The current reference range fortiagabine is 0.020.2 mg/L (0.050.53 mmole/L).1 However,Uthman et al167 showed in their patient group that the reductionin the frequency of complex partial seizures was related to serumtiagabine concentrations with the best seizure control occurringat trough concentrations above 0.04 mg/L (0.11 mmole/L),whereas concentrations of 0.4 mg/L (1.06 mmole/L) were asso-ciated with central nervous system toxicity.

Saliva TDM for TiagabineIt is not known whether tiagabine is secreted into saliva

and if it is whether concentrations are similar to or reectthose in serum. However, because serum concentrations are12 orders of magnitude less than most other AEDs and witha protein binding of 96%, measurement of the free nonpro-tein bound concentration will be analytically challenging.

Topiramate

Clinical IndicationsTopiramate is licensed for the monotherapy treatment

of generalized tonicclonic seizures and partial seizures with



or without secondarily generalization in adults and childrenaged 6 years and above; also as adjunctive therapy for adultsand children aged 2 years and above. It is also licensed for theadjunctive treatment of seizures associated with LennoxGastautsyndrome and for primary generalized tonicclonic seizures.Topiramate is also licensed for the treatment of migraine andis available as formulations of tablet and a sprinkle capsule.

Pharmacokinetic CharacteristicsTopiramate is rapidly absorbed after oral ingestion with

a Tmax of 24 hours and bioavailability is .80%. Its pharma-cokinetics is linear and protein binding is 15%.168 Approxi-mately 50% of a dose of topiramate is excreted unchanged bythe kidneys with the remainder being metabolized in the liver,primarily by yet to be identied CYP isoenzymes, to formseveral oxidative metabolites some of which are conju-gated.169,170 The serum elimination half-life of topiramate inadults is 2030 hours and the drug is subject to many drugdrug pharmacokinetic interactions consequent to the fact thatit is a weak inducer of hepatic metabolism; furthermore, itsown metabolism can be enhanced by hepatic enzyme inducersand some drugs also inhibit topiramate metabolism.103 Conse-quently, there are large differences between individuals in thedose to serum concentration relationship. The current referencerange for topiramate in serum is 520 mg/L (1559 mmole/L).1

Saliva TDM for TopiramateThere has been one study investigating the distribution

of topiramate into saliva and the relationship between salivaryand serum topiramate concentration (both total and free) inadults and children with epilepsy.78 Topiramate distributesinto saliva such that the salivary concentration is similar tothe total concentration in serum with saliva/serum total con-centration ratios ranging 0.631.13 (mean 0.90 6 0.12). Theconcentrations of topiramate in saliva are signicantly corre-lated (r2 = 0.97) with the total in serum, which enables salivato be used as an alternative matrix for topiramate TDM.

Valproate

Clinical IndicationsValproate is licensed for monotherapy or adjunctive

treatment of any form of epilepsy in patients of any age. It isalso licensed for the treatment of migraine and bipolardisorder. Valproate is available in a variety of formulationsincluding, enteric-coated tablets, chewable tablets, capsules,sustained-release tablets and microspheres, modied releasegranules, a liquid oral solution, a syrup, and a solution forintravenous injection.

Pharmacokinetic CharacteristicsValproic acid is rapidly absorbed after oral ingestion;

however, Tmax values are formulation dependent and variable(17 hours). Bioavailability is generally.90% but is 8%20%lower for extended release formulations. The serum proteinbinding of valproate is normally 90% but is concentration depen-dent and saturable, which causes the pharmacokinetics of thedrug to be nonlinear.171 Valproic acid is extensively metabolizedin the liver and metabolism is complex in that it involves multiplemetabolic pathways including O-glucuronidation, b-oxidation,

v-oxidation, hydroxylation, ketone formation, and desatura-tion. In excess of 25 metabolites have been identied but val-proic acid glucuronide and 3-oxo-valproic acid are by far themost abundant. The serum elimination half-life of valproic acidin adults is 1216 hours,172 which reects the interindividualvariation in clearance and it is subject to many drugdrugpharmacokinetic interactions.103 Valproic acid is a potent inhib-itor of hepatic metabolism and furthermore its own metabolismcan be both induced and inhibited; consequently, there arelarge differences between individuals in the dose to serumconcentration relationship. The current reference range for totalvalproic acid in serum is 50100 mg/L (346693 mmole/L),1

but the measurement of free nonprotein bound valproic acidconcentration is more clinically useful because of the largevariability in protein binding.

Saliva TDM for ValproateThe physicochemical properties of valproic acid (ie,

a weak acid with a pKa of 4.9), and the pH gradient betweenserum and saliva result in only small quantities of valproicacid passing into saliva; furthermore, concentrations areerratic. Stimulation by citric acid did not increase the lowsaliva valproic acid concentrations, and the correlationbetween saliva and serum valproic acid concentrations (bothtotal and free) was poor.54,173 Because salivary valproic acidconcentrations correlate poorly with total and free serum val-proic acid concentrations,54,95,173176 saliva cannot be used asan alternative matrix for valproic acid TDM.

Vigabatrin

Clinical IndicationsVigabatrin is licensed for the adjunctive treatment of

partial seizures with and without secondary generalization. Itis also licensed for the monotherapy treatment of infantilespasms (West Syndrome). Vigabatrin is available as formu-lations of a tablet and a sachet (containing granules) only.

Pharmacokinetic CharacteristicsVigabatrin is rapidly absorbed after oral ingestion with

a Tmax of 12 hours and a bioavailability of 60%80%. Itspharmacokinetics is linear, and it is not protein bound.177

Vigabatrin is not metabolized being cleared entirely by renalexcretion with a serum elimination half-life of 58 hours inadults.178 Vigabatrin is not subject to drugdrug pharmacoki-netic interactions.103 It acts by elevating brain g-aminobutyric acid (GABA) concentrations. GABA is the principalinhibitory neurotransmitter in the brain, and vigabatrin actsthrough a mechanism whereby it selectively and irreversiblybinds to GABA-transaminase (GABA-T) thus blocking theenzyme responsible for GABA metabolism and increasingGABA concentrations.179 GABA concentrations remain ele-vated for some time after vigabatrin can be detected in serumbecause GABA metabolism depends on resynthesis ofGABA-T.180 Because of its mechanism of action, TDM ofvigabatrin is generally considered not to be very helpfuland a wide range of trough serum concentrations have beenfound in patients successfully treated with the drug. However,TDM may be useful in patients with renal impairment, to



ascertain compliance, where malabsorption is suspected andin cases of suspected overdose. The current reference rangefor vigabatrin in serum is 0.836 mg/L (6279 mmole/L).1

Saliva TDM for VigabatrinAfter administration of a single oral dose of vigabatrin

to 6 healthy volunteers, salivary vigabatrin concentration wasapproximately 10% of that in serum.79 Thus, salivary vigaba-trin concentrations do not reect the free nonprotein boundconcentrations in serum. There are no data as to the correla-tion between salivary and serum vigabatrin concentrations,and therefore, it is uncertain at present as to whether salivacould be used as an alternative matrix for vigabatrin TDM.

Zonisamide

Clinical IndicationsZonisamide is licensed for the adjunctive treatment of

partial seizures with and without secondary generalization inadults. Zonisamide is available only as formulations of a capsule.

Pharmacokinetic CharacteristicsZonisamide is rapidly absorbed after oral ingestion with

a Tmax of 25 hours and bioavailability of .90%. Its pharma-cokinetics is linear, and protein binding is 40%; also, the drughas a high afnity, low capacity-binding site on erythrocytes.181

Zonisamide is extensively metabolized in the liver, primarily byacetylation, to form N-acetyl zonisamide and reduction catalyzedby CYP3A4, to form 2-sulfamoylacetylphenol, which is sub-sequently glucuronidated.182 The serum half-life of zonisamidein adults is 5070 hours so that interindividual clearance is vari-able.183185 Zonisamide is subject to some drugdrug pharmaco-kinetic interactions consequent to the fact that its metabolism canbe induced or inhibited103; consequently, there are large differ-ences between individuals in the dose to serum concentrationrelationship. The current reference range for zonisamide inserum is 1040 mg/L (47188 mmole/L).1

Saliva TDM for ZonisamideZonisamide distributes into saliva, but limited informa-

tion is available relating salivary concentrations to those inserum. The drug seems to be actively secreted into salivabecause concentrations are reported to be higher than the freeconcentration in serum.80 More recently, Jones et al186 includedsaliva samples from patients prescribed zonisamide in a mailingstability study, and although they demonstrated that the drugwas stable in saliva, the authors did not compare the individualresults with matching serum specimens.

INDICATIONS FOR SALIVA TDMThere are numerous indications for saliva AED TDM,

and these are similar to those for serum AED TDM1:

Dose Optimization on the InitiallyPrescribed AED

Measurement of saliva concentrations of the initiallyprescribed AED can be of particular value whenever the besttherapeutic response has been achieved in an individual andmaintained for a sufcient period of time to be condent that

dosage has been optimized (ideally this would be seizurefreedom but for many patients it would entail optimumseizure control with minimal and tolerated adverse effects). Inthis situation, measuring the saliva AED concentration at astandardized sampling time will allow identication of theindividual therapeutic concentration, which will be usefulto interpret the clinical situation should a change in responseoccur during further follow-up.1 The advantage of the indi-vidual therapeutic concentration approach is that it does notrely on xed reference ranges and can be applied to anyAED, including all the new AEDs for some of which refer-ence ranges have yet to be clearly dened. Ideally, whenestablishing the individual therapeutic concentration, 2 sepa-rate determinations obtained at intervals of 24 months willbe preferable to a single concentration measurement, becausethis allows an estimate of any variability.

Uncontrolled SeizuresKnowledge of the individual therapeutic concentration

can greatly enhance management in patients who developbreakthrough seizures after a prolonged period of seizurecontrol. For example, after a breakthrough seizure has occurred,if the salivary concentration is much lower than the previouslydetermined individual therapeutic concentration it would sug-gest either suboptimal compliance or a clinically importantchange in the pharmacokinetics of the AED.51,187 In the settingwhereby seizures persist despite an apparently adequate dosageof an appropriate AED, saliva AED concentration monitoringwould be useful to identify potential causes of therapeutic fail-ure which may result from poor compliance (typically charac-terized by variable saliva concentrations, which increase aftersupervised drug intake) or from poor absorption, fast metabo-lism or drug interactions (typically characterized by low salivaAED concentrations).

In a series of 95 subjects with uncontrolled seizures thatwere admitted to an emergency department, saliva AED con-centrations were able to reveal that noncompliance was respon-sible for the seizures in 31% of patients.188 Zysset et al189 ina series of 13 children with epilepsy observed that salivary phe-nytoin concentrations agreed well to that of clinical judgmentabout compliance and Herkes and Eadie138 were able to identifynoncompliance as a contributor to uncontrolled seizures in theirpatients by salivary TDM.

Suspected ToxicityIf toxic symptoms are suspected in a patient, saliva

AED concentrations can aid in conrming a diagnosis of drugtoxicity. For example, Cohen et al120 showed a signicantcorrelation between saliva concentrations and toxicity forphenytoin, whereas Hamilton et al190 showed that carbama-zepine saliva conc

Documents

Saliva Aed Tdm 2013