Kenneth W. Sommerville MD
Medical Director, Neuroscience, Department of Marketed Products, Clinical Research, Abbott Laboratories, Abbott Park, Illinois
Tiagabine (Gabitril) is a nipecotic acid derivative that appears to suppress seizures by inhibiting the uptake of γ-aminobutyric acid (GABA) by presynaptic neurons and glial cells. Nipecotic acid, which is unable to cross the blood-brain barrier, was identified as a GABA uptake inhibitor in vitro; subsequently, tiagabine was synthesized by adding a lipophilic anchor to the amino acid nitrogen of nipecotic acid. Tiagabine readily crosses the blood-brain barrier, and it demonstrates inhibition of GABA uptake by neurons and glial cells similar to nipecotic acid in vitro (1). Tiagabine produces a significant increase in extracellular GABA levels in rat brains at doses that inhibit pentylenetetrazol-induced tonic seizures (2).
In comparison with other antiepileptic drugs (AEDs), tiagabine is well tolerated over a wide range of doses. It appears to have minimal or no effect on the pharmacokinetics of other drugs, in part because it is used therapeutically at low concentrations. The pharmacokinetics of tiagabine are, however, affected by other AEDs. The interactions of tiagabine are limited and predictable, possibly because the drug is active at low concentrations and has a well-defined mechanism of action. In this chapter, studies regarding the effects of other drugs on tiagabine and the effects of tiagabine on other drugs are reviewed.
To understand the interactions between tiagabine and other drugs, it is helpful to review the basics of tiagabine pharmacokinetics. Tiagabine is a drug of high potency and relatively low toxicity: in vitro studies have shown it to be the most potent inhibitor of GABA uptake yet identified (1), and therapeutic benefits are observed in the dose range of 24 to 80 mg/day when the agent is coadministered with enzyme-inducing AEDs (3). The median lethal dosage (ED50) of tiagabine in rats and mice is at least 50-fold higher than the anticonvulsant ED50 (4). In humans, the concentrations of tiagabine associated with clinical use are in the range of nanograms per milliliter. These low concentrations may be a principal reason that the agent has little or no effect on other drugs.
The bioavailability of tiagabine is >90% whether it is administered as tablets, capsules, or oral solution (5). Tiagabine is rapidly absorbed over its complete therapeutic dose range, with the maximum concentration (Cmax) typically occurring from 0.5 to 1.5 hours of administration under fasting conditions. It also demonstrates linear pharmacokinetics over the dose range of clinical use; both the dose-adjusted Cmax and the dose-adjusted area under the plasma concentration-time curve (AUC) are independent of dose. The plasma concentration-time curve of tiagabine typically includes secondary peaks, a finding indicating the possibility of enterohepatic recycling; and the harmonic mean half-life is roughly 7 hours (6).
Tiagabine appears to be metabolized principally through the action of the 3A subfamily of cytochrome P450 (CYP) enzymes, a conclusion drawn from in vitro experiments involving incubation of tiagabine with hepatic microsomes and studies employing selective CYP3A inhibitors (7). In plasma, tiagabine is highly protein bound (>95%). Tiagabine does not induce hepatic enzymes, as suggested by its lack of effect on antipyrine metabolism (6). No active metabolites of tiagabine have been identified (7). Although food intake slows the rate of tiagabine absorption, it does not appear to affect the extent of absorption; the AUC0—∞ in the fed state is similar to that in the fasting state (5).
EFFECTS OF OTHER ANTIEPILEPTIC DRUGS ON TIAGABINE
By far the most important effect of other AEDs on tiagabine pharmacokinetics is the increased rate of tiagabine metabolism and clearance in the presence of drugs known to induce hepatic metabolism. Because tiagabine is often given in combination therapy or as an add-on to existing AED therapy, this effect is important for dosage recommendations.
The pharmacokinetics of tiagabine was evaluated in an early single-dose trial of safety and tolerability in the presence of other AEDs (8). In this study, a single 8-mg dose of tiagabine was administered to four groups of patients who had been receiving long-term treatment with valproate alone, carbamazepine and phenytoin, carbamazepine and primidone, or carbamazepine and vigabatrin. Previously evaluated normal volunteers, not taking any AEDs, served as historical controls. Carbamazepine, phenytoin, and primidone are known to induce hepatic enzymes, whereas valproate does not. As anticipated, when antipyrine was administered to evaluate enzyme induction, the groups taking known hepatic enzyme-inducing drugs demonstrated elevated antipyrine clearance and decreased half-life. After administration of tiagabine, the mean AUCs for the historical controls and for the valproate group were 754 and 908 ng-hr/mL, respectively. However, the mean AUCs for the other three (enzyme-induced) groups were significantly lower (260, 437, 239 ng·hr/mL, respectively; p < .05 for each group). This study suggested the possibility that, in the presence of enzyme-inducing AEDs, the dosage of tiagabine may need to be increased, to provide the same AUC and presumably the same therapeutic effect as when the drug is taken alone or with non-enzyme-inducing AEDs (8).
Another study evaluated the pharmacokinetics of tiagabine when used as add-on therapy in a stable regimen involving one to three enzyme-inducing AEDs, including carbamazepine, phenytoin, phenobarbital, and primidone (9). Patients in this study were enrolled in one of two longterm trials of tiagabine, as add-on therapy to other AEDs. This study confirmed the linear pharmacokinetics of tiagabine in the presence of other AEDs; when groups taking 40, 56, or 80 mg/day of tiagabine were compared, the dose-adjusted Cmax, trough concentration (Cmin), and AUC0-6 hr values were independent of dose. The harmonic mean half-life ranged from 3.8 to 4.9 hours in the three dosage groups (individual values ranged from 1.7 to 8.8 hours). When compared with the harmonic mean half-life of 7.1 hours in a group of historical control volunteers who were not taking enzyme-inducing AEDs, these results again demonstrated that tiagabine metabolism and clearance is increased in the presence of enzyme-inducing AEDs (9).
The increased clearance, lower AUC, and shorter half-life of tiagabine in the presence of enzyme-inducing AEDs was further substantiated by a population pharmacokinetic analysis across a large number of patients in two long-term clinical trials of epilepsy. One trial included only patients with partial seizures; the other included patients with any kind of epilepsy. Testing was conducted on a total of 2,147 plasma samples from 511 patients (age range, 11 to 77 years); tiagabine dosage ranged from 2 to 80 mg/day, and the sampling times relative to the previous dose of tiagabine were distributed across a wide range. This study included patients using tiagabine as add-on therapy to enzyme-inducing AEDs (carbamazepine, phenytoin, primidone, or barbiturates), tiagabine as add-on therapy to noninducing AEDs (valproate or clonazepam), and tiagabine monotherapy. The data were analyzed using the NONMEM program (10), which permits the analysis of limited data from large numbers of patients in clinical settings to yield an estimate for oral clearance value (CL/F; intravenous clearance over bioavailability), as well as other pharmacokinetic parameters. In this study, the central oral clearance value CL/F was 21.4 L/hr in enzyme-induced patients, compared with 12.8 L/hr in noninduced patients (Figure 73.1). The increased clearance value for tiagabine in enzyme-induced patients is consistent with other studies demonstrating accelerated clearance in such patients. The effect of more than one enzyme-inducing AED on tiagabine pharmacokinetics was not additive: clearance values in patients undergoing stable therapy with more than one inducing AED were similar to those in which tiagabine was used as add-on therapy to a single inducing AED. This study showed that the tiagabine dose should be adjusted (usually increased) when a new enzyme-inducing AED is added to an otherwise noninducing AED regimen that includes tiagabine. The dose of tiagabine should also be adjusted (usually reduced) when inducing AEDs are discontinued as concomitant therapy (11).
After administration, tiagabine in plasma is highly protein bound (>95%), primarily to albumin and α1-acid glycoprotein. If its pharmacokinetic properties are determined in part by the nature and extent of this binding, they could potentially be affected by drugs that displace tiagabine or otherwise affect its binding. An extensive study of various AEDs and other drugs, however, showed that tiagabine is tightly bound, and little or no change occurs in tiagabine binding in the presence of other drugs. Tiagabine is not displaced by phenytoin, carbamazepine, or phenobarbital, nor is there clinically important displacement of these drugs in plasma. Tiagabine binding is also unaffected by various other drugs that were studied in vitro, including propranolol, verapamil, chlorpromazine, amitriptyline, imipramine, warfarin, ibuprofen, digitoxin, furosemide, tolbutamide, and haloperidol (12). Both salicylic acid and naproxen displace tiagabine from protein, but this effect is unlikely to be clinically relevant.
Valproate has been found to affect the binding of tiagabine to plasma proteins, by reducing the bound fraction by a small, but statistically significant, amount—from 96.3% to 94.8% (12). Tiagabine, however, has no effect on valproate binding. Moreover, clinical studies involving concomitant administration of tiagabine and valproate have demonstrated that the clearance values and AUC of tiagabine remain similar to those observed when tiagabine is administered alone (8,11). Therefore, the effect of the minimally reduced binding level with the addition of valproate is clinically unimportant.
The metabolism of tiagabine by the CYP3A family of cytochrome P450 enzymes raises the concern that its pharmacokinetics may be affected by known inhibitors of this
enzyme family (13). Both cimetidine and erythromycin affect cytochrome P450 activity through interaction with its heme moiety; cimetidine itself forms a tight complex with the heme iron, whereas a metabolite of erythromycin is the heme-binding species. In a multiple-dose crossover study, the plasma Cmin and AUC of tiagabine were slightly increased (~5%) when cimetidine was concurrently administered with tiagabine to healthy volunteers, but the changes were not judged to be clinically important (14).
FIGURE 73.1. Distribution of post hoc clearance (CL/F) estimates in patients included in the NONMEM analysis. (From Samara EE, Gustavson LE, El-Shourbagy T, et al. Population analysis of the pharmacokinetics of tiagabine in patients with epilepsy. Epilepsia 1998;39:868-873, with permission.)
In a two-period crossover study, erythromycin (500 mg twice daily) was administered to 13 healthy volunteers concurrently with tiagabine (4 mg twice daily) during one period, and placebo was given with tiagabine during the other period. Erythromycin treatment was initiated 12 hours before the initial tiagabine dose and was continued until 24 hours after the final tiagabine dose. The AUC, Cmax, and half-life of tiagabine were identical in the presence and absence of erythromycin. The only change noted was a decrease in the time to reach maximum plasma level (Tmax) of tiagabine in the presence of erythromycin (15). These studies indicate that the effects of drugs known to bind to cytochrome P450 on tiagabine metabolism are slight and are of little or no clinical importance.
EFFECT OF TIAGABINE ON OTHER DRUGS
Because tiagabine does not induce hepatic enzymes and is present in small quantities in serum at clinically relevant doses, its effect on pharmacokinetics of other AEDs administered concurrently is expected to be minimal. This has proved to be the case in pharmacokinetic studies. Two single-center open-label studies examined the potential effects of the addition of tiagabine on the pharmacokinetics and safety of carbamazepine or phenytoin at steady state. In adult patients with seizures controlled by a stable fixed dose of either phenytoin (n = 12) or carbamazepine (n = 12), tiagabine was administered in a starting dose of 8 mg/day for 3 days. The doses were then escalated at 3-day intervals to 16, 32, and 48 mg/day (final dose level). Serial blood samples were obtained on day 1 (before tiagabine initiation) and on day 18 (after the final tiagabine dose). Mean Cmax, Tmax, Cmin, and AUC0-τvalues remained unchanged in the presence of tiagabine, both for patients receiving carbamazepine and for those receiving phenytoin (16). The mean plasma concentration-time profiles in the presence and absence of tiagabine are shown for carbamazepine (Figure 73.2A) and for phenytoin (Figure 73.2B).
FIGURE 73.2. A: Mean carbamazepine (CBZ) and carbamazepine epoxide (CBZE) plasma concentrations after treatment with CBZ alone or in combination with tiagabine (TBG).B: Mean phenytoin (PHT) plasma concentration-time profiles after treatment with phenytoin alone or in combination with TGB. (A and B, from Gustavson LE, Cato A, Boellner SW, et al. Lack of pharmacokinetic drug interactions between tiagabine and carbamazepine or phenytoin. Am J Ther 1998;5:9-16, with permission.)
Another study, of similar design, evaluated the interaction between tiagabine and valproate in patients with seizures controlled by a fixed stable dose of valproate. In this study, the tiagabine dose was started at 8 mg/day, escalating at 3-day intervals to 24 mg/day. Blood samples were obtained on day 1 (before tiagabine initiation) and day 14 (after the final tiagabine dose). The serum concentrations of both valproate and tiagabine were determined; tiagabine pharmacokinetics, as expected, were similar in the presence of valproate (a noninducer) to those observed when tiagabine was administered alone. The mean Cmax and AUC0-τ values for valproate were approximately 10% to 12% lower when the drug was coadministered with tiagabine than when it was given alone (Figure 73.3). Much of this variability originated from two patients (out of 12 total) whose valproate values declined >30% during the tiagabine regimen. These patients had a history of missed tiagabine doses, and they may have also missed one or more valproate doses as well; when their values were omitted, the reduction in valproate values fell to ~6%. A possible alternative explanation for these patients' results involved the timing of valproate administration and the use of Depakote, an enteric-coated formulation of valproate. In any case, the small reduction in valproate concentrations was thought to be of limited clinical importance, because valproate has a broad dose range in epilepsy (17).
The effects of tiagabine on other (non-AED) drugs also appear to be minimal. In separate multiple-dose studies, each involving 10 to 23 healthy volunteers, tiagabine did not significantly alter the pharmacokinetics of several other drugs. Twelve healthy male volunteers received open-label (12 mg/day) tiagabine, alone or with digoxin (0.5 mg/day on day 1, then 0.25 mg/day on days 2 to 9), in a two-period crossover design. Pharmacokinetic parameters measured on day 9 of each period showed no significant differences between the groups (18). Both (R)-and (S)-warfarin parameters were measured in a parallel-group, double-blind study in 25 healthy male subjects. After achieving a prothrombin time of 14 to 18 seconds for 7 days, tiagabine 12 mg/day or placebo was added for
5 days, and pharmacokinetic parameters were measured. There were no significant differences between the pharmacokinetics or pharmacodynamics of the (R)- and (S)- enantiomers with or without the addition of tiagabine. In another double-blind study, 14 healthy male subjects were given theophylline 200 mg every 6 hours for 5 days and a single dose of tiagabine 10 mg or placebo on day 5. Pharmacokinetic parameters of theophylline showed no significant differences with the addition of either tiagabine or placebo. Tiagabine also did not alter the sedative effect or cognitive impairment caused by triazolam or ethanol, although the duration of triazolam-induced effects was observed to be slightly prolonged (14).
FIGURE 73.3. Mean valproate plasma concentration-time profiles after treatment with valproate (VPA) alone or in combination with tigabine (TGB). (From Gustavson LE, Sommerville KW, Boellner SW, et al. Lack of a clinically significant pharmacokinetic drug interaction between tiagabine and valproate. Am J Ther 1998;5:73-79, with permission.)
The interaction of tiagabine with oral contraceptives was studied in 10 healthy female volunteers taking combination formulations (30 µg ethinyl estradiol with 150 µg of either levonorgestrel or desogestrel) at least 6 months before the study and over two complete pill-taking cycles (28 days) during the study. After a control period (cycle 1, days 1 to 23), 8 mg/day tiagabine was administered for 13 days (cycle 1, days 24 to 28 and cycle 2, days 1 to 7), followed by an evaluation period (cycle 2, days 8 to 28). Plasma hormonal levels were measured throughout both cycles, and bleeding patterns were recorded. The progesterone level on days 19 to 21 of each cycle was used as the indicator of nonovulation; all patients had levels that remained lower than the established threshold level of < 13 nmol/L, signaling nonovulation. No induction of hepatic enzymes was observed, based on urinary levels of 6β-hydroxycortisol, an indicator for induction. Two patients experienced breakthrough bleeding, although both remained well below the accepted progesterone threshold for nonovulation. In general, tiagabine was well tolerated, and it demonstrated little or no interaction in combination with oral contraceptives (19).
Because AEDs are usually administered over long periods (often lifelong), it is important to evaluate their interactions with other substances likely to be consumed. The potential interactions between tiagabine and ethanol were investigated using tests selected from the Cognitive Drug Research Computerized Cognitive Assessment System. The selected tests included immediate and delayed word recall and delayed word recognition, digit vigilance (speed and accuracy), choice reaction time, visual tracking, and body sway. Twenty healthy volunteers underwent a double-blind, placebo-controlled, two-period crossover study. The study involved an escalating dose of tiagabine or placebo over 9 days; on the final day, they consumed a standardized ethanol dose (0.7 g/kg for men, 0.6 g/kg for women). Serial cognitive tests were performed, and blood samples obtained, before and after ethanol administration. The pharmacokinetics of tiagabine and ethanol was not significantly affected by concurrent administration of both. As expected, ethanol caused a significant impairment in performance on the most sensitive assay, the speed and accuracy of digit vigilance. However, concurrent administration of tiagabine did not significantly affect performance on this or any other test, a finding indicating that tiagabine does not potentiate the depressant and cognitive effects of ethanol (20).
The general neuropsychological effects of tiagabine were investigated in 162 adult patients with difficult-to-control seizures (at least six complex partial seizures during the 8 weeks before the initial screening visit). All patients were receiving a stable regimen including at least one enzyme-inducing AED. Patients either received placebo or were stabilized at one of three tiagabine dose levels (16, 32, or 56 mg/day) that were titrated from a starting dose of 8 mg/day. A battery of tests, measuring both abilities (including various cognitive skills) and adjustment (including mood, quality of life, and psychosocial variables), was conducted at baseline and after the stable dose period. The evaluations of abilities and adjustment showed no significant differences between placebo and tiagabine groups at the p < .01 level and four differences at the p < .05 level. The authors noted that six statistically significant (p < .05) differences would have been expected on the basis of chance alone. They concluded that no evidence of a clear drug effect exists; however, the data do not rule out possible small effects that are unlikely to be clinically important. A reduction in complex seizure frequency of >50% was observed in some (16 of 71) of the patients receiving either 32 or 56 mg/day of tiagabine, and this may have been expected to result in significant improvements in measures of adjustment. However, even among patients experiencing such marked improvement, the typical patient would still have had a seizure every 2 weeks on average. This pattern would not alter many of the life-changing attributes of poorly controlled epilepsy (e.g., the inability to drive or the ability to take on new psychosocial roles). Therefore, any improvement may not have been substantial enough to make an important difference in the quality of life (21).
The dose-related effects of tiagabine monotherapy on cognition and mood were studied in 123 adult patients with uncontrolled partial seizures. All patients were treated with only one marketed AED at the time of study entry. The study consisted of an 8-week baseline period, a 6-week withdrawal and titration period, and a 12-week fixed-dose period. At the end of baseline, patients were randomized to 6 mg/day tiagabine monotherapy (n = 66) or to 36 mg/day tiagabine monotherapy (n = 57) under double-blind conditions. Baseline AEDs were tapered beginning at day 15 and were discontinued entirely by 5 weeks after the first dose of the study drug. The 12-week fixed-dose period began after the 6-week withdrawal and titration period. Results showed
that few changes in either cognition (abilities) or mood (adjustment) were noted when all patients were considered as a single group. However, analysis of both dose and attainment of tiagabine monotherapy showed that patients able to attain tiagabine monotherapy had improvements compared with the groups at the same dose that failed to achieve monotherapy. The group receiving 6 mg/day improved primarily in the areas of adjustment and mood, and the group receiving 36 mg/day improved in the area of abilities. Failure to attain tiagabine monotherapy was associated with worsening on tests of mood and adjustment in patients receiving 36 mg/day tiagabine. It was suggested that worsening of mood could have been avoided if a slower titration schedule had been used, because there was difficulty in tolerating the 36-mg dose. The investigators concluded that attainment of tiagabine monotherapy at high or low dose was associated with the improvement of varying degree on neuropsychological tests. The withdrawal of the baseline AED may have contributed to improvement in both dose groups (22).
Tiagabine is an AED with few documented drug-drug interactions. The pharmacokinetic parameters of tiagabine include rapid absorption and virtually complete bioavailability, linear kinetics, and metabolism by the CYP3A family of cytochrome P450 enzymes. Tiagabine does not induce hepatic microsomal enzymes, although its clearance is increased by hepatic enzyme-inducing drugs.
Among other drugs that may be frequently combined with tiagabine, the most clinically important interaction is the increased clearance of tiagabine when it is administered concurrently with hepatic enzyme-inducing AEDs. Because of this interaction, the initial dosage of tiagabine should be considered carefully when it is added to an existing AED regimen. In addition, when tiagabine is part of a stable AED regimen, the dosage should be adjusted on the addition or removal of enzyme-inducing AEDs.
Tiagabine has not been found to have significant effects on the pharmacokinetics of other AEDs or other commonly used drugs, including oral contraceptives and drugs commonly used by elderly patients (digoxin, warfarin). Tiagabine does not potentiate the effects of ethanol. Tiagabine appears to have little or no effect on standardized measures of psychological abilities and adjustment when it is added to other AEDs in patients with refractory partial seizures. Improvement in neuropsychological testing has been noted when tiagabine monotherapy is attained after conversion from other AEDs.
I gratefully acknowledge the review of this manuscript by Linda Gustavson, PhD.