Seizure (2004) 13, 574—581

The pharmacokinetic inter-relationship oftiagabine in blood, cerebrospinal fluidand brain extracellular fluid (frontalcortex and hippocampus)

Xiaolan Wang, Neville Ratnaraj, Philip N. Patsalos*

Pharmacology and Therapeutics Unit, Department of Clinical and Experimental Epilepsy, Institute ofNeurology, Queen Square, London WC1N 3BG, UK

KEYWORDSTiagabine;Pharmacokinetics;Blood;Cerebrospinal fluid;extracellular fluid;Hippocampus;Frontal cortex

Summary Purpose: Tiagabine is a unique antiepileptic drug with a novel mecha-nism of action. Whilst some limited data are available as to the peripheral bloodpharmacokinetics of tiagabine, data regarding the kinetics of tiagabine in the cen-tral brain compartment are very limited. We therefore sought to investigate serum,cerebrospinal fluid (CSF) and frontal cortex and hippocampal extracellular fluid (ECF)kinetic inter-relationship of tiagabine in a freely moving rat model. Methods: Adultmale rats were implanted with either a jugular vein catheter and a cisterna magnacatheter for blood and CSF sampling, respectively, or a blood catheter and a micro-dialysis probe in the hippocampus and frontal cortex (for ECF sampling). Tiagabine wasadministered intraperitoneal (i.p.) at 20 or 40mg/kg and blood, CSF and ECF were col-lected at timed intervals for themeasurement of tiagabine concentrations by high per-formance liquid chromatography. Results: Tiagabine concentrations in blood and CSFrose linearly and dose-dependently and time to maximum concentration (Tmax) was15 and 29min, respectively. Mean CSF/serum tiagabine concentration ratios (range,0.008—0.01) were much smaller than the mean free/total tiagabine concentrationratios in serum (0.045± 0.003). Entry of tiagabine into brain ECF (frontal cortex andhippocampus) was rapid with Tmax values of 31—46min. Distribution of tiagabine inbrain was not brain region specific with values in the frontal cortex and hippocampusbeing indistinguishable. Whilst elimination from CSF was comparable to that of serum,half-life (t1/2) values in ECF were three times longer. Conclusions: Tiagabine is asso-ciated with linear kinetic characteristics and with rapid brain penetration. However,CSF concentrations are not reflective of free non-protein-bound concentrations inserum. The observation that tiagabine elimination from the brain is threefold slowerthan that seen in blood, may explain as to the relatively long duration of action oftiagabine.© 2004 BEA Trading Ltd. Published by Elsevier Ltd. All rights reserved.

*Corresponding author. Tel.: +44-20-7837-3611x3830;fax: +44-20-7278-5616.

E-mail address: P.Patsalos@ion.ucl.ac.uk (P.N. Patsalos).

1059-1311/$30 — see front matter © 2004 BEA Trading Ltd. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.seizure.2004.01.007

The pharmacokinetic of tiagabine 575

Introduction

Tiagabine ([R-(−)-N-(4,4-di-(3-methylthien-2-yl)-but-3-enyl) nipecotic acid] hydrochloride) is aunique antiepileptic drug (AED) with a novel mech-anism of action in that it selectively and specif-ically inhibits the uptake of �-aminobutyric acid(GABA) into astrocytes and neurones, by the trans-porter GAT-1, and thus increases the extracellularconcentration of GABA in the brain. Tiagabine iseffective in the management of partial seizureswith or without secondary generalisation, and islicensed for use as adjunctive therapy in patientsrefractory to available first-line AEDs.1—4

Tiagabine is effective in a variety of seizuremodels including amygdala-kindled seizures,5 au-diogenic seizures6 and convulsions associated withstatus epilepticus induced in cobalt-lesioned rats bythe administration of homocysteine thiolactone.7

The latter observation raises the possibility that itmay be useful in the management of status epilep-ticus in patients. However, from a pharmacokineticaspect, drugs used in the management of statusepilepticus need to be able to rapidly and effi-ciently penetrate the blood—brain barrier and tohave an elimination half-life that does not requirefrequent dosing.Whilst data on the blood pharmacokinetics of

tiagabine in man are abundant,8 such data arescarce in the rat.9 Furthermore, data regardingthe kinetics of tiagabine in the brain are very lim-ited indeed; one study has reported on whole braintiagabine concentrations at 1 h after intraperito-nial administration of tiagabine.10 Tiagabine con-centrations increased dose-dependently and werehigher in the dorsal cortex compared to those mea-sured in the cerebellum. Some additional informa-tion on the central brain kinetics of tiagabine canbe deduced indirectly via observation of centralGABA changes. Thus, tiagabine dose-dependentlyincreased GABA extracellular fluid (ECF) concentra-tions two- to threefold in the globus pallidus, ven-tral pallidum and in the substantia nigra of rats.11

However, tiagabine had no effect on cerebrospinalfluid (CSF) GABA concentrations after 2 days oftiagabine administration (50—200mg/kg/day).12 Amicrodialysis study of a single patient with epilepsy,reported an ∼50% increase in GABA ECF concentra-tions at 1 h after a single dose of oral tiagabine in-gestion and the increase was sustained for severalhours.13

The present study sought to ascertain theperipheral blood pharmacokinetic and centralbrain (CSF and ECF) kinetic inter-relationship oftiagabine using a well established and validated ratmodel.14—16

Methods

Animals

Male Sprague—Dawley rats weighing 250—300 g (A.Tuck & Son Ltd., Battlesbridge, Essex, UK) weregroup-housed in contiguous cages and fed ad libitumon normal laboratory diet (SDS R and M number 1expanded, Scientific Dietary Services, Witham, Es-sex, UK) and water. A 12-h light—dark cycle (lightson 06:00 h) and ambient temperature of 25 ◦C weremaintained. All the rats were allowed to acclimatiseto their new environment for at least 3 days priorto experimentation. All animal procedures strictlyfollowed Home Office regulations and were per-formed under the Animal (Scientific Procedures) Act1986.

Surgical procedures, blood, CSF andmicrodialysis sampling

Rats were anaesthetised with 2% halothane (MerialAnimal Health Ltd., Dublin, Ireland) inhalation. Twosurgical procedures were undertaken. In the firstgroup of rats, catheters were implanted in the rightjugular vein for blood sampling and in the cisternamagna for CSF sampling as previously described.14

For the second group of rats the procedure in-volved the implantation of a catheter in the rightjugular vein for blood sampling and a microdialy-sis probe in both the hippocampus (from bregma5.6mm posterior, 5mm lateral, 8.2mm ventral)and the frontal cortex (from bregma 2.5mm ante-rior, 1.5mm lateral, 5.5mm ventral) for monitoringof the ECF. Stereotaxic placement of the probeswas according to the atlas of Paxinos and Watson.17

Post-surgery rats were housed singly in perspexcages.Two days later, when animals had fully recov-

ered, the microdialysis probes and the CSF andblood catheters were checked for patency. Animalswith blood and CSF catheters were investigated asfollows: 20 or 40mg/kg tiagabine was administeredby intraperitoneal (i.p.) injection. Blood (200�l)and CSF (20�l) were sampled via the implantedcatheters at timed intervals over a 7-h period (15,30, 45, 60, 90, 120, 150, 180, 240, 300, 360, 390and 420min). In order to prevent the developmentof hypovolemia, an equivalent volume of hep-arinized saline was administered after each bloodsampling. Blood samples were collected in 0.5mlpolyethylene tubes (Treff AG, Switzerland), vortexmixed and centrifuged at 10,000 rpm for 5min (Ab-bott micro-centrifuge, Abbott Diagnostics, Maid-enhead, UK). Supernatant serum and CSF were

576 X. Wang et al.

stored frozen (−70 ◦C) until analysis for tiagabinecontent.Animals with blood catheters and microdialy-

sis probes were investigated as follows: ArtificialCSF (CSF composition (in mM): NaCl 125, KCl 2.5,MgCl2 1.18 and CaCl2·1.26) was perfused throughthe probes at 2�l/min. A baseline sample of blood(200�l) and three baseline dialysate samples (20�l)were taken before drug administration. The ratswere then injected i.p. with 40mg/kg tiagabineand venous blood samples (200�l) were withdrawnat 10, 20, 30, 40, 50, 60, 75, 90, 105 and 120min.Dialysate samples (20�l) were collected at 10-minintervals for the first 60min and at 15-min inter-vals for the subsequent 60-min period and samplingtime was corrected for dead-space volume time.Blood and microdialysate samples were collectedin 0.5ml polyethylene tubes (Treff AG). Bloodsamples were prepared as described above and su-pernatant serum and microdialysate were storedfrozen (−70 ◦C) until analysis for tiagabine content.

Microdialysis probe construction andin vitro recovery

Concentric microdialysis probes with filtral 12(Hospal, Rugby, UK) dialysis membrane (4mm long,200�m diameter) and internal vitreous silica tub-ing (SGE, Milton Keynes, UK) were prepared aspreviously described. In vitro probe recovery wasdetermined by placing each microdialysis probeinto a beaker containing a solution of 200 nmol/ltiagabine, constituted in artificial CSF, and thenperfusing with artificial CSF at 37 ◦C with a flowrate of 2�l/min. Samples (40�l) were collectedevery 20min for 120min and stored at −70 ◦C untilanalysis for tiagabine content.

Measurement of tiagabine concentration

Tiagabine concentrations in serum, CSF andmicrodialysates were measured by high perfor-mance liquid chromatography (HPLC). The HPLCsystem comprised: a Gilson 307 pump, a Gilson 234autoinjector, ESA coulochem electrochemical de-tector and an analytical cell (model 5011). A Hyper-sil BDS-C18, 3�m, 125mm× 3mm column (HewlettPackard, Stockport, Cheshire, UK) and LiChro-sphere select B 4× 4 (5�m) pre-column (HewlettPackard) were used. Chromatograms were collatedusing Unipoint System Software (Version 1.71).The lower limit of quantification for tiagabine was0.5 nmol/l (CV, 6%). The lower limit of detectionwas 0.1 nmol/l. The procedure used for the de-termination of the free non-protein-bound serum

tiagabine concentration was the same as for to-tal tiagabine concentrations except that sampleswere first filtered through an Amicon CentrifreeMicropartition System (Amicon, Stonehouse, UK)using a Sigma 2K15 centrifuge with a temperaturesetting of 25 ◦C.

Kinetic and statistical analysis

Pharmacokinetic parameters were computed (Mi-crosoft Excel) based on a one-compartment modelfor CSF and microdialysate and a two-compartmentmodel for blood with first order elimination. Thefollowing pharmacokinetic parameters were deter-mined: time to maximal concentration (Tmax), max-imal concentration (Cmax), terminal eliminationhalf-life (t1/2) and area under the concentrationversus time curve (AUC). The AUC was calculatedby linear trapezoid summation and extrapolation toinfinite time as appropriate. The Ke was estimatedby least-square regression analysis of the termi-nal log-linear portion of the plasma concentrationversus time profile. The (t1/2) was calculated bydividing the natural log of 2 by Ke. Cmax and Tmaxwere read directly from the concentration versustime curves.Non-parametric Mann—Whitney tests were used

to ascertain significance and a P value of 0.05 wasconsidered statistically significant.

Results

Blood pharmacokinetics

Tiagabine pharmacokinetic parameters in serumafter 20 and 40mg/kg tiagabine administration areshown in Table 1. Tiagabine concentrations rosedose-dependently and Cmax values were achieved atthe time of first sampling (15min post-dose). Con-centrations subsequently declined rapidly and thedecline was exponential (Fig. 1). Both Cmax and AUCvalues increased linearly and dose-dependently.Tiagabine t1/2 values were 55± 2.3min and50± 2.6min after 20 and 40mg/kg administration,respectively, and were not statistically different(P>0.05, Mann—Whitney test). At 240min and atsubsequent time points tiagabine concentrationswere just quantifiable and were of the order of150 nmol/l.The free non-protein-bound tiagabine concentra-

tion was determined in serum. At 15min post-dose(the time of first sample), the tiagabine free frac-tionwas 1.4%. However, subsequent samples (n = 8)showed a larger free fraction value (4.5± 0.3%) andwere both time and concentration independent.

The pharmacokinetic of tiagabine 577

Table 1 Pharmacokinetic parameters of tiagabine in serum, CSF and brain ECF after tiagabine administration at20 and 40mg/kg (n = 5—7).

Tmax (min) Cmax (nmol/l) AUC (nmol h/l) t1/2 (min)

SerumTiagabine (20mg/kg) 15 ± 0 14905 ± 1733 9439 ± 909 55 ± 2.3Tiagabine (40mg/kg) 16 ± 0.3 27754 ± 2000 25147 ± 3491 50 ± 2.6

CSFTiagabine (20mg/kg) 28 ± 0.7 49 ± 6.7 51 ± 7.6 40 ± 2.6Tiagabine (40mg/kg) 32 ± 0.9 95 ± 5.4 138 ± 13 64 ± 2.7

ECFTiagabine (40mg/kg)Frontal cortex 41 ± 5 35 ± 1 154 ± 25 174 ± 32Hippocampus 34 ± 3 38 ± 2 134 ± 10 133 ± 9

All values are mean± S.E.M. Tmax: time to maximal concentration; Cmax: maximal concentration; AUC: area underthe concentration vs. time curve; t1/2: terminal elimination half-life; CSF: cerebrospinal fluid; ECF: extracellularfluid.

0

5000

10000

15000

20000

25000

30000

0 60 120 180 240 300 360 420

Time (min)

TG

B c

on

cen

trat

ion

in s

eru

m (

nm

ol/L

)

Figure 1 Tiagabine (TGB) concentration vs. time profiles in serum after 20mg/kg (�) and 40mg/kg (�) TGBadministration. Values are mean± S.E.M. (n = 5—7).

Neuropharmacokinetics

CSFThe kinetic constants after 20 and 40mg/kgtiagabine administration in CSF are shown inTable 1. After drug administration, tiagabinerapidly and readily penetrated into the CSF com-partment and tiagabine was detectable in CSF atthe time of first sampling (15min; Fig. 2). MeanTmax values varied between 28 and 32min and weredose-independent. However, compared to serum,CSF Tmax values were double the value (30min ver-sus 15min). Cmax and AUC values increased linearlyand dose-dependently. In contrast to serum CSFt1/2 values were significantly longer after 40mg/kg

tiagabine administration (64± 2.7min) comparedto when 20mg/kg tiagabine was administered(40± 2.6min; P < 0.01, Mann—Whitney test).Fig. 3 shows the tiagabine concentration ratio of

CSF/serum over time after 40mg/kg tiagabine ad-ministration. There was a tendency towards equi-libration (as determined by a constant CSF/serumtiagabine concentration ratio) by 30min post-tiagabine administration. Rapid equilibration wasalso observed after administration of 20mg/kgtiagabine (data not shown). Mean CSF/serum ra-tios were 0.008± 0.0006 after 20mg/kg tiagabineadministration and 0.01± 0.0009 after 40mg/kgtiagabine administration. These values are not sta-tistically different (P>0.05, Mann—Whitney test).

578 X. Wang et al.

0

20

40

60

80

100

120

0 60 120 180 240 300

Time (min)

TG

B c

on

cen

trat

ion

in C

SF

(n

mo

l/L)

Figure 2 Tiagabine (TGB) concentration vs. time profiles in CSF after 20mg/kg (�) and 40mg/kg (�) TGB adminis-tration. Values are mean± S.E.M. (n = 5—7).

Extracellular fluid kineticsThe mean± S.E.M in vitro recovery for tiagabinefor all the microdialysis probes was 10± 0.3% at adialysate flow rate of 2�l/min. The artificial CSFsolution used contained 200 nmol/l tiagabine andthese recovery data were used to adjust the in vivoECF concentration data.The pharmacokinetic parameters of tiagabine in

brain ECF frontal cortex and hippocampus after40mg/kg tiagabine administration are shown inTable 1. The corresponding tiagabine concentrationversus time profiles are shown in Fig. 4. After drugadministration, blood—brain barrier penetrationwas rapid and tiagabine was detectable at the timeof first sampling (10min post-dose). Mean Tmaxvalues for frontal cortex and hippocampus were41± 5min and 34± 3min, respectively. AUC values

0.00

0.01

0.02

0 60 120 180 240 300 360 420

Time (min)

CS

F /

seru

m T

GB

co

nce

ntr

atio

n r

atio

Figure 3 CSF/serum tiagabine (TGB) concentration ratios vs. time after 40mg/kg TGB administration. Values aremean± S.E.M. (n = 5).

were comparable for frontal cortex and hippocam-pus (154± 25 nmol h/l versus 134± 10 nmol h/l,respectively). Overall, there were no statisticallysignificant difference in tiagabine kinetic parame-ters between brain frontal cortex and hippocam-pus (P>0.05, Mann—Whitney test). However, thehalf-life values in ECF frontal cortex and hippocam-pus were significantly longer (up to three times)than that seen in the CSF and in blood.Fig. 5 shows the tiagabine concentration ratio

of ECF/serum for frontal cortex and hippocampusover time after 40mg/kg tiagabine administra-tion. In contrast to CSF, there was no tendencytowards equilibration (as determined by a risingECF/serum tiagabine concentration ratio) dur-ing the study period of up to 120min. However,the mean ECF/serum ratios for frontal cortex

The pharmacokinetic of tiagabine 579

0

5

10

15

20

25

30

35

40

45

0 15 30 45 60 75 90 105 120

Time (min)

TG

B c

on

cen

trat

ion

in b

rain

EC

F (

nm

ol/L

)

Figure 4 Tiagabine (TGB) concentration vs. time profiles in frontal cortex (�) and hippocampal (�) ECF after40mg/kg TGB administration. Values are mean± S.E.M. (n = 5—7).

0

0.002

0.004

0.006

0.008

0.01

0 30 60 90 120

Time (min)

EC

F /

seru

m T

GB

co

nce

ntr

atio

n r

atio

Figure 5 ECF/serum tiagabine (TGB) concentration ratios vs. time profiles in frontal cortex (�) and hippocampal(�) after 40mg/kg TGB administration. Values are mean± S.E.M. (n = 5—7).

and hippocampus were indistinguishable (P>0.05,Mann—Whitney test).

Discussion

This is the first study to investigate the temporalpharmacokinetic inter-relationship of tiagabine inblood, CSF and brain ECF. The major findings of thisstudy are: (1) the pharmacokinetics of tiagabine inserum are linear and dose-dependent; (2) tiagabineis substantially protein bound (95%) in serum; (3)the elimination of tiagabine from serum is rapid(t1/2 values, 50—55min) and independent of dose;(4) tiagabine rapidly and readily penetrated the

blood—brain barrier and was detectable within min-utes in both CSF and brain ECF; (5) although the CSFpharmacokinetics of tiagabine paralleled that seenin serum, CSF concentrations did not reflect freedrug concentrations in serum; (6) whilst tiagabinet1/2 values were comparable in CSF and serum, val-ues were three times larger in frontal cortex andhippocampal ECF; (7) tiagabine distribution in brainfrontal cortex and hippocampal ECF were indistin-guishable.Although there is a lack of published studies

describing the kinetics of tiagabine in the rat,the sparse data reported by Suzdak and Jansen9

are generally in agreement with those describedin the present report. After oral ingestion of an

580 X. Wang et al.

unspecified tiagabine dose, tiagabine Tmax valueswere ∼30min;9 slightly longer, as might be ex-pected, compared to after i.p. administration inthe present study (Table 1). Tiagabine serum pro-tein binding was also comparable (92—93% versus95—96%, this study). That in the present study ECFtiagabine t1/2 values were three times longer thanthat observed in the blood compartment (Table 1),corroborates the indirect data presented by Suz-dak and Jansen.9 It was noted that although after[14C]tiagabine administration (dose and route ofadministration were not reported) the decreasein tissue concentrations of [14C]tiagabine (liver,kidney and lung) followed the decrease in serumconcentrations, the brain/plasma ratio remainedconstant. This was interpreted by the authors as in-dicating a slower elimination of tiagabine from thebrain or that equilibration between serum and brainis slow. Indeed, the data presented in the presentstudy would suggest that both these processesmight pertain (Table 1 and Fig. 5). Furthermore,the CSF t1/2 values reported in the present studywere significantly larger at the higher tiagabinedose (40mg/kg; 64± 2.7min) when comparedto values obtained at the lower dose (20mg/kg;40± 2.6min). These data corroborate the obser-vation that elimination of tiagabine from the brainwas dose-dependent.9 Lastly, the observation thatthe distribution of tiagabine in the frontal cortexand hippocampus is similar is in agreement with thein vitro tissue binding studies with [3H]tiagabine.18

The rate of penetration of drugs, including AEDs,into CSF and brain is considered to be determinedessentially by three physiochemical characteristics,namely degree of ionisation, lipid solubility and de-gree of serum protein binding. In relation to AEDs,lipid solubility appears to be the major determinantof rate of entry into the CSF compartment.19 Be-cause tiagabine has high lipid solubility it would beexpected to gain rapid entry into the brain and in-deed this was the case as tiagabine was detectablein both CSF and frontal cortex and hippocampal ECFat the time of first sampling (10—15min post-dose).Furthermore, the rate of entry into the CSF was notdose-dependent, as there was no difference in Tmaxvalues after different tiagabine dose administration(20 and 40mg/kg). Thus, tiagabine transportationacross the blood—CSF barrier in the choroid plexuswas not rate-limiting at the concentration rangesachieved. This characteristic may very well apply totransportation of tiagabine across the blood—brainbarrier but since the kinetics of tiagabine in frontalcortex and hippocampal ECF were determined af-ter only one tiagabine dose (40mg/kg) this can-not be definitively conclude from the presentdata.

Free (non-protein-bound) drug concentrationsin serum are often used as a reflection of CNSdrug concentrations. In the present study, meanCSF/serum tiagabine concentration ratios (range,0.008—0.01) are much smaller than the meanfree/total tiagabine concentration ratios in serum(0.045± 0.003). These data suggest that freetiagabine serum concentrations do not accuratelyreflect CSF tiagabine concentrations. This is in con-trast to other AEDs (e.g. carbamazepine, pheny-toin, primidone, phenobarbitone and lamotrigine)studied in this animal model where CSF concentra-tions do reflect free drug concentrations in serumand emphasises the importance of studying thecentral neuropharmacokinetics of a drug at the siteof drug action.14,16,20,21

That tiagabine elimination from the brain is muchslower than that that occurs in serum could be ofsignificant clinical importance since it would sug-gest that the pharmacological effect of tiagabine ismuch longer than that that would be suggested byits blood pharmacokinetic profile. Thus, the shorttiagabine t1/2 in serum, which is generally per-ceived as a disadvantage for the clinical therapeu-tics of tiagabine,16 may actually not be as relevantif indeed the data observed in the present studycould be extrapolated to man. A further consid-eration in relation to the difference between thecentral and peripheral kinetics of tiagabine wouldbe that if the dosing strategy for tiagabine is basedon its serum kinetics, tiagabine would accumulatein the brain. Thus, if non-convulsive status epilep-ticus, that has been reported in patients treatedwith tiagabine,22—24 is the consequence of GABAer-gic hyperfunction, as has been suggested,25 then itcould be the consequence of tiagabine accumula-tion in the brain.The tiagabine doses used in the present study (20

and 40mg/kg) are higher than the ED50 for tiagabinein animal seizure models26 or indeed doses thatare used clinically.4 Because tiagabine is highly pro-tein bound (95—96%), only a very small proportion(4—5%) of the tiagabine concentration achieved inblood passes the blood—brain barrier to enter theCSF and ECF compartments. Consequently, dose se-lection was based on the rationale and the needto achieve quantifiable tiagabine concentrations inCSF and ECF. That the doses used in the presentstudy do not impede on the value of the data col-lected, is evidenced by the fact that the overallconclusions reached are similar to those describedby Suzdak and Jansen.9

Finally, the regional distribution of tiagabine inthe brain needs some comment. The present studyclearly shows that the distribution of tiagabinein frontal cortex and hippocampal ECF is almost

The pharmacokinetic of tiagabine 581

identical and the data corroborate data from invitro tissue binding studies with [3H]tiagabine.18

GABA transporters (designated GAT-1, GAT-2,GAT-3 and BGT-1 according to their pharmacolog-ical sensitivity)27 have a differential anatomicaldistribution throughout the nervous system.28,29

As GAT-1 is primarily located in the cerebral cor-tex and hippocampus and as tiagabine (through itsspecificity for GAT-1) is equally distributed in thesebrain regions, it would be expected the pharmaco-logical action of tiagabine would be brain regionspecific.In conclusion, this study has shown that tiagabine

has linear peripheral and central kinetics and withrapid brain and CSF penetration although CSF con-centrations do not reflect free drug concentrationsin serum. Its distribution in brain cerebral cortexand hippocampus would suggest it is not region spe-cific and its elimination from the brain greatly out-lasts that seen in blood.

Acknowledgements

The authors would like to thank Norvo Nordisk forthe provision of pure tiagabine.

References

1. Sachdeo RC, Leroy RF, Krauss GL, et al. Tiagabine-therapyfor complex partial seizures: a dose-frequency study. ArchNeurol 1997;54:595—601.

2. Kalviainen R, Brodie MJ, Duncan J, Chadwick D, Edwards D,Lyby K. A double blind placebo-controlled trial of tiagabinegiven three times daily as add-on therapy for refractoryseizures. Epilepsy Res 1998;30:31—40.

3. Uthman BM, Rowan AJ, Ahmann PA, et al. Tiagabine for com-plex partial seizures: a randomized, add-on, dose-responsetrial. Arch Neurol 1998;55:56—62.

4. Gaitatzis A, Patsalos PN, Sander JWAS. The clinical efficacyof tiagabine. Rev Contemp Pharmacother 2002;12:235—50.

5. Dalby NO, Nielsen EB. Tiagabine exerts ananti-epileptogenic effect in amygdala kindling epileptoge-nesis in the rat. Neurosci Lett 1997;229:135—7.

6. Faingold CL, Randall ME, Anderson CA. Blockade of GABAuptake with tiagabine inhibits audiogenic seizures and re-duces neuronal firing in the inferior colliculus of the genet-ically epilepsy-prone rat. Exp Neurol 1994;126:225—32.

7. Walton NY, Gunawan S, Treiman DM. Treatment of exper-imental status epilepticus with the GABA uptake inhibitortiagabine. Epilepsy Res 1994;19:237—44.

8. Wang X, Patsalos PN. The pharmacokinetic profile oftiagabine. Rev Contemp Pharmacother 2002;12:225—33.

9. Suzdak PD, Jansen JA. A review of the preclinical pharma-cology of tiagabine: a potent and selective anticonvulsantGABA uptake inhibitor. Epilepsia 1995;36:612—26.

10. Sills GJ, Patsalos PN, Butler E, Forrest G, Ratnaraj N, BrodieMJ. Visual field constriction: accumulation of vigabatrin butnot tiagabine in the retina. Neurology 2001;57:196—200.

11. Fink-Jensen A, Suzdak PD, Swedberg MD, Judge ME, HansenL, Nielsen PG. The gamma-aminobutyric acid (GABA) up-

take inhibitor, tiagabine, increases extracellular brain lev-els of GABA in awake rats. Eur J Pharmacol 1992;220:197—201.

12. Halonen T, Nissinen J, Jansen JA, Pitkanen A. Tiagabineprevents seizures neuronal damage and memory impair-ment in experimental status epileptics. Eur J Pharmacol1996;299:69—81.

13. During M, Mattson R, Scheyer R. The effect of tiagabine HClon extracellular GABA levels in the human hippocampus.Epilepsia 1992;33(Suppl 3):83.

14. Patsalos PN, Alavijeh MS, Semba J, Lolin YI. A freely mov-ing and behaving rat model for the chronic and simultane-ous study of drug pharmacokinetics (blood) and neurophar-macokinetics (cerebrospinal fluid): haematological and bio-chemical characterisation and kinetic evaluation using car-bamazepine. J Pharmacol Toxicol Methods 1992;28:21—8.

15. Tong X, Patsalos PN. A microdialysis study of the novelantiepileptic drug levetiracetam: extracellular pharmacoki-netics and effect on taurine in rat brain. Br J Pharmacol2001;133:867—74.

16. Wang X, Patsalos PN. A comparison of central brain (cere-brospinal and extracellular fluids) and peripheral blood ki-netics of phenytoin after intravenous phenytoin or fos-phenytoin. Seizure 2003;12:330—6.

17. Paxinos G, Watson C. The rat brain in stereotaxic coordi-nates. 2nd ed. San Diego: Academic Press; 1986.

18. Suzdak PD, Foged C, Anderson KE. Quantitative autora-diographic characterisation of the binding of [3H]tiagabine(NNC 05-328) to the GABA uptake carrier. Brain Res1994;647:231—41.

19. Loscher W, Frey HH. Kinetics of penetration of com-mon antiepileptic drugs into cerebrospinal fluid. Epilepsia1984;25:346—52.

20. Nagaki S, Ratnaraj N, Patsalos PN. Blood and cere-brospinal fluid pharmacokinetics of primidone and its pri-mary pharmacologically active metabolites, phenobarbitaland phenylethylmalonamide in the rat. Eur J Drug MetabPharmacokinet 1999;24:255—64.

21. Walker MC, Tong X, Perry H, Alavijeh MS, Patsalos PN. Com-parison of serum, cerebrospinal fluid and brain extracellu-lar fluid pharmacokinetics of lamotrigine. Br J Pharmacol2000;130:242—8.

22. Schapel G, Chadwick D. Tiagabine and non-convulsive statusepilepticus. Seizure 1996;5:153—6.

23. Eckardt KM, Steinhoff BJ. Nonconvulsive status epilepticusin two patients receiving tiagabine treatment. Epilepsia1998;39:671—4.

24. Ettinger AB, Bernal OG, Andriola MR, et al. Two cases of non-convulsive status epilepticus in association with tiagabinetherapy. Epilepsia 1999;40:1159—62.

25. Solomon GE, Labar D. Hypothesis that tiagabine-inducedNCSE is associated with GABAergic hyperfunction, withGABA(B) receptors playing a critical role, is supported bya case of generalised NCSE induced by baclofen. Epilepsia1998;39:1383.

26. Walker MC. The mechanism of action of tiagabine. RevContemp Pharmacother 2002;12:213—23.

27. Borden LA, Smith KE, Hartig PR, Branchek TA, Weinshank RL.Molecular heterogeneity of the �-aminobutiric acid (GABA)transport system. J Biol Chem 1992;267:21098—104.

28. Durkin MM, Smith KE, Borden LA, Weinshank RL, BranchekTA, Gustafson EL. Localisation of messenger RNAs encodingthree GABA transporters in rat brain: an in situ hybridizationstudy. Mol Brain Res 1995;33:7—21.

29. Ribak CE, Tong WMY, Brecha NC. The GABA plasma mem-brane transporters, GAT-1 and GAT-3, display different dis-tribution in the rat. J Comp Neurol 1995;367:595—606.