Address correspondence and reprint requests to Dr. S.J. Czuczwar at Department of Pathophysiology, Medical University, Jaczewskiego 8, 20-090 Lublin, Poland. E-mail: firstname.lastname@example.org
Summary: Purpose: The study investigated the types of interactions between lamotrigine (LTG) and first-generation antiepileptic drugs (AEDs) or topiramate (TPM) with isobolographic analysis.
Methods: Anticonvulsant and adverse-effect profiles of combinations of LTG with other AEDs, at fixed ratios of 1:3, 1:1, and 3:1, were evaluated in the maximal electroshock (MES)-induced seizures and the chimney test (motor performance) in mice, which allowed the determination of benefit indices (BIs) for individual combinations.
Results: Combinations of LTG with TPM or valproate (VPA), at fixed ratios of 1:1, were significantly supraadditive (synergistic) in the MES test and, simultaneously, subadditive (antagonistic) in the chimney test, showing the best profile for AED combinations. In contrast, combinations between LTG and carbamazepine (CBZ), in terms of antiseizure protection against MES, were subadditive (antagonistic) and additive in the chimney test, resulting in unfavorable AED combinations. Moreover, the combination of LTG with phenobarbital (PB), at a fixed ratio of 1:1, despite synergy in the MES test, also was synergistic in the chimney test, resulting in a modest BI for AED combination. LTG combined with phenytoin was additive in both the MES and chimney tests in mice. The remaining combinations, at fixed ratios not mentioned earlier, also showed an average BI for AED combinations. Furthermore, LTG combined with all studied AEDs did not affect long-term memory in mice. None of the AEDs influenced the free plasma level of LTG, whereas LTG slightly reduced the free plasma concentration of PB.
Conclusions: Interactions between LTG and TPM or LTG and VPA at a fixed ratio of 1:1 might be profitable from a preclinical point of view, displaying the most optimal BI.
No doubt the proper selection of antiepileptic drugs (AEDs), with respect to seizure types and epilepsy syndromes, is indispensable to provide the efficacious seizure control in epilepsy patients. Monotherapy as the “gold standard” in epilepsy treatment usually offers a satisfactory seizure medication in ∼70% of patients (1–3). However, in spite of advances in the understanding of pathophysiologic processes involved in seizure initiation and propagation, as well as significant insight into molecular mechanisms of antiseizure activity of commonly available AEDs, some patients (30%) do not adequately respond to the AEDs in monotherapy. In such cases, when the three sequential monotherapy approaches with the current frontline AEDs fail, a rational polytherapy, based on theoretic presumptions and preclinical (animal) studies, should be used to enhance effectiveness in seizure control (4–8).
From a clinical point of view, pharmacologic therapy in epileptic patients should provide complete seizure control and simultaneously produce favorable adverse-effect profiles. In the 1990s, eight newer AEDs were approved and introduced into clinical practice, primarily as adjunctive drugs added to the standard therapy with conventional AEDs. It became evident that several AED combinations may offer some improvements in patients' quality of living through a significant reduction of seizures and/or favorable adverse-effect profiles (9,10). Among the most efficacious duotherapies in humans are some drug combinations containing lamotrigine (LTG), a second-generation AED.
The mechanism of action of LTG [3,5-diamino-6-(2,3-dichlorophenyl)-1,2,4-triazine] is very similar to that of phenytoin (PHT) and carbamazepine (CBZ) (i.e., the drug acts at voltage-dependent Na+ channels to decrease the presynaptic release of the excitatory neurotransmitter glutamate) (11). LTG, binding to the inactivated form of Na+ channels, limits the sustained repetitive firing of neurons, leaving normal synaptic conduction unaffected (12,13). The drug selectively blocks veratradine-evoked but not K+-evoked release of endogenous glutamate (14). Moreover, LTG decreases voltage-gated Ca2+ currents (15), and this effect also contributes to a decrease in glutamate release and thereby to the anticonvulsant action of LTG. In experimental studies, LTG is active in amygdala-kindled rats and maximal electroshock (MES)-induced seizures in mice but is ineffective against pentylenetetrazol (PTZ)-induced clonus (16,17). However, the preclinical profile of LTG differs significantly from that of PHT and CBZ in that LTG is effective in the lh/lh mice [a model of absence seizures (18)], whereas PHT and CBZ can even worsen absence seizures (19,20). The preclinical profile of LTG supports its clinical utility against partial, generalized tonic–clonic, and absence seizures (21–23).
No presumptions have been experimentally proven about the choice of combinations of the available AEDs with LTG to enhance the effectiveness of treatment regimen in patients with drug-resistant epilepsy. Theoretically, effective AED combinations were empirically evaluated in patients with intractable seizures; however, such evaluations were often accompanied with deleterious adverse-effect reactions (24). According to Deckers et al. (8), an isobolographic analysis (as the optimal method for detecting synergy, additivity, or antagonism among AEDs in animal models of epilepsy) may indicate which of these AED combinations are expected to provide the patients with adequate control of refractory seizures.
This preclinical study investigated the types of interactions between LTG and conventional AEDs [CBZ, PHT, valproate (VPA), and phenobarbital (PB)] or topiramate (TPM) to establish the rationale for specific AED combinations to control refractory epileptic attacks.
MATERIALS AND METHODS
The experiments were performed on Swiss male mice, purchased from a licensed dealer. The animals were kept under standardized laboratory conditions with free access to food and tap water and a natural light–dark cycle for 7 days before the experiments. The experimental groups consisted of eight to 10 animals, weighing 20–26 g. All experiments were carried out between 9:00 a.m. and 3:00 p.m. The experimental procedures listed were approved by the Local Bioethical Committee of Lublin Medical University (license 191/2001/207/01).
The following AEDs were used in this study: LTG (Lamictal; Glaxo Wellcome, Kent, U.K.), valproate magnesium (VPA) kindly donated by ICN-Polfa S.A. (Rzeszów, Poland), CBZ (a gift from Polfa, Starogard, Poland), PHT (Polfa, Warsaw, Poland), PB (Polfa, Cracow, Poland), and TPM (Topamax; Cilag AG, Schaffhausen, Switzerland). LTG, CBZ, PB, PHT, and TPM were suspended in a 1% solution of Tween 80 (Sigma, St. Louis, MO, U.S.A.), whereas VPA was dissolved in sterile saline. The studied AEDs were injected intraperitoneally (i.p.) in a volume of 10 ml/kg; VPA and CBZ, 30 min; LTG, PB, and TPM, 60 min; and PHT, 120 min before electroconvulsions and behavioral tests. Routes of i.p. administration and pretreatment times before testing of the AEDs were based on information about their biologic activity from the literature (25) and confirmed in our previous experiments (26), as well as pilot experiments in case of LTG.
Electroconvulsions were produced with a Hugo Sachs generator (Rodent Shocker, type 221; Freiburg, Germany) with the use of auricular electrodes (50 Hz, 25 mA). The stimulus duration was 0.2 s. The end point was the tonic extension of the hind limbs. The protective activity of AEDs was determined as their ability to protect 50% of mice against the MES-induced tonic hind limb extension, and expressed as respective median effective dose (ED50) values. To evaluate an ED50 value, at least four groups of mice, after receiving progressive doses of an AED, were challenged with MES-induced convulsions. A dose–response curve for each AED was subsequently constructed on the basis of a percentage of animals protected against the tonic convulsions.
The chimney test, elaborated and described in detail by Boissier et al. (27), was used to assess the range of doses of the studied AEDs producing motor impairment in animals. In this test, animals had to climb backward up a plastic tube (3 cm inner diameter, 25 cm length). Motor impairment was indicated as the inability of mice to climb backward up the tube within 60 s. Adverse neurotoxic effects of the studied AEDs were expressed as their median toxic dose (TD50) values, representing the doses at which the AEDs impair motor coordination in 50% of the animals. At least four groups of mice, consisting of eight to 10 animals, were used to evaluate each TD50 value. A dose–response curve for each AED was subsequently calculated on the basis of a percentage of mice showing motor deficits.
Step-through passive-avoidance task
The animals were given AEDs separately or in combinations on the first day before training. The time before the training session (after drug administration) was identical to that for the MES test. Subsequently the animals were placed in an illuminated box (10 × 13 × 15 cm) connected to a large dark box (25 × 20 × 15) equipped with an electric grid floor. Entrance to the dark box was punished by an adequate electric foot shock (0.6 mA for 2 s). The animals that did not enter the dark compartment were excluded from the experiment. On the next day (24 h later), the pretrained animals (without treatment) were put again into the illuminated box and observed for ≤180 s. Mice that avoided the dark compartment for 180 s were considered to remember the task. The time at which the mice entered the dark box was noted, and subsequently, the medians with 25th and 75th percentiles were calculated. The step-through passive-avoidance task gives information about ability to acquire the task (learning) and to recall the task (retrieval). Therefore it may be regarded as a measure of long-term memory (28).
The isobolographic analysis is based on a comparison of drug doses determined to be equieffective. The experimental (Dmix) and the theoretic additive (Dadd) ED50 or TD50 values were determined from the dose–response curves of combined drugs (29). ED50 is defined as a dose of a drug protecting 50% of the animals against MES-induced seizures. TD50 is defined as a dose of an AED causing impairment of motor coordination in 50% of the animals. Dmix is an experimentally determined total dose of the mixture of two component drugs, which were administered in the fixed-ratio combination sufficient for a 50% protective effect (for ED50 mix) or a 50% neurotoxic effect (for TD50 mix). Conversely, Dadd represents a total additive dose of two drugs (calculated from the line of additivity), theoretically providing 50% protection against seizures (for ED50 add) or 50% impairment of motor coordination in mice (for TD50 add). The respective 95% confidence limits of Dmix were calculated according to Litchfield and Wilcoxon (30), and these of Dadd according to Tallarida and Murray (31), and subsequently transformed to standard errors of the mean (SEM), according to the procedure described in our previous study (26). To estimate the types of interactions, three fixed-dose ratios of the drugs were examined as follows: 1:3, 1:1, and 3:1 (for all tested combinations) in the MES and chimney tests in mice.
In the present study, only interactions distinctly diverse from additivity were graphically presented as the isobolograms. To visualize the types of interactions between LTG and AEDs studied, the isoboles were drawn by plotting the points reflecting the respective doses of LTG (on the X-axis), and doses of an AED on the Y-axis. The straight line connecting either ED50 or TD50 values for the two tested drugs, administered alone, against MES and in the chimney test, represents the theoretic isobole for additivity. If experimentally determined data points, reflecting the combinations of various fixed ratios, fall significantly below the additivity line, the two component drugs act synergistically. Conversely, antagonism may be recognized if these points are localized above the additive isobole.
Immunofluorescence estimation of the free plasma concentrations of antiepileptic drugs
AEDs were analyzed quantitatively in plasma of mice at the time scheduled for the MES test. The animals were administered an AED + vehicle or a combination of LTG with the respective AED tested. The fixed-drug ratio combination (LTG:an AED) for estimating the free plasma levels of AEDs was chosen as 1:1 for all AEDs. Free plasma levels of VPA, CBZ, PB, PHT, and TPM administered alone or in combination with LTG were measured by immunofluorescence, with the use of an Abbott TDx automatic analyzer (Abbott, Irving, TX, U.S.A.), according to the procedure described in our previous studies (26,32). Reagents for the assays of conventional AEDs were purchased from Abbott Laboratories (Abbott) and that for TPM from Oxis International, Inc. (Portland, OR, U.S.A.). Plasma levels of AEDs were expressed in micrograms per milliliter of plasma, as means ± SD of at least eight determinations.
Chromatographic determination of lamotrigine plasma concentration
LTG was analyzed quantitatively in plasma of animals at times scheduled for the MES-induced seizures in mice. The animals were administered LTG + vehicle or a combination of LTG with AEDs, at the fixed ratio of 1:1. Free plasma levels of LTG were evaluated according to the procedure described in our earlier study (32). Plasma concentrations of LTG were expressed in micrograms per milliliter of plasma, as means ± SD of at least eight determinations.
Both ED50 and TD50 values (with 95% confidence limits) were calculated by computer probit analysis (30) and subsequently transformed into SEM (26). Statistical analysis of drug interactions was performed by the use of Student's t test to evaluate the difference between experimental (Dmix) and theoretical additive (Dadd) ED50 and TD50 values. Free plasma levels of AEDs (VPA, CBZ, PB, PHT, and TPM) alone or in combination with LTG, as well as the free plasma concentrations of LTG after administering AEDs, were statistically analyzed with Student's t test. Moreover, the Kruskal–Wallis nonparametric analysis of variance (ANOVA) test followed by post hoc Dunn's test was used for analyzing data from the step-through passive-avoidance task.
1Protective index (PI) was calculated by dividing a given TD50, obtained in the chimney test, by the respective ED50 determined in the MES test. The PI is considered a satisfactory margin of safety between AED doses and doses of AEDs exerting sedative, ataxic, or other neurotoxic side effects (33).
2Benefit index (BI) was defined as a quotient of PImix and PIadd of respective fixed-ratio combinations, obtained directly from the isobolographic analysis. PImix is a protective index experimentally determined, and PIadd is a protective index theoretically calculated from the lines of additivity in the MES and chimney tests. BI unequivocally estimates advantages of the combination of two drugs applied in various fixed-ratio combinations. Moreover, BI provides the rationale for combining the respective AEDs in clinical practice if its value is >1.3, whereas BI <0.7 indicates unfavorable combinations of AEDs (26).
Isobolographic analysis of the protection offered by lamotrigine combined with studied antiepileptic drugs in maximal electroshock–induced seizures in mice
All studied AEDs (LTG, VPA, CBZ, PB, PHT, and TPM) produced dose-dependent anticonvulsant effects against MES-induced seizures in mice. The ED50 values for the drugs administered alone are presented in Table 1. Based on ED50 values for separate AEDs, the theoretically additive total drug mixtures (ED50 add values) were calculated for three fixed ratios (1:3, 1:1, and 3:1). Subsequently, the experimental ED50 mix values were determined for the same fixed-ratio combinations in the MES test (Table 2). The isobolographic analysis showed supraadditive (synergistic) interactions between LTG and VPA (at the dose ratios of 1:3 and 1:1; Fig. 1A), LTG and PB (1:3, 1:1; Fig. 2A), and LTG and TPM (1:1; Fig. 3A). With isobolography, a subadditive (antagonistic) interaction was observed for the combination of LTG with CBZ at fixed-dose ratios of 1:1 and 3:1 (Fig. 4A). The remaining combinations between LTG and AEDs tested were evaluated as purely additive in the MES test (Table 2).
Table 1. Influence of LTG and some AEDs studied on their protective effect against MES = induced seizures and on the impairment of motor coordination in the chimney test in mice
Data presented as median ED50 or TD50 values (with 95% confidence limits in parentheses) for the respective AEDs evaluated in the MES and chimney tests.
AED, antiepileptic drug; ED50, median effective dose; TD50, median toxic dose; MES, maximal electroshock; PI, protective index was calculated as the quotient of TD50 and ED50 values, for every AED tested, to assess the margin of safety of AED therapy; LTG, lamotrigine; PHT, phenytoin; VPA, valproate; PB, phenobarbital; CBZ, carbamazepine; TPM, topiramate.
Table 2. Isobolographic analysis of interaction between LTG and some AEDs in the MES and chimney tests in mice
Results presented as the ED50 or TD50 values (in mg/kg) ± SEM, experimentally determined from isobolographic analysis (ED50 mix or TD50 mix) or theoretically calculated from the additivity line (ED50 add or TD50 add).
F, fixed ratio of drug dose combinations; PImix and PIadd, protective indices denoted as the quotient of the respective TD50s and ED50s values for the respective fixed-ratio drug combination; BI, benefit index, as a quotient of PImix and PIadd, unequivocally classifies drug interactions as advantageous or unfavorable for clinical practice; a value >1.3 indicates advantageous combinations, worth recommendation to clinical practice, whereas a value <0.7 characterized combinations that should be avoided in the antiseizure treatment. Statistical analysis of data was performed by using Student's t test.
a p < 0.01; b p < 0.05; c p < 0.001 vs. the respective ED50 add or TD50 add group.
LTG + VPA
136.4 ± 7.9a
178.2 ± 8.6
271.7 ± 8.8b
301.9 ± 8.3
96.8 ± 5.5a
120.3 ± 6.0
234.8 ± 7.5b
210.8 ± 7.0
67.9 ± 3.0
62.3 ± 3.4
117.3 ± 6.1
119.8 ± 5.7
LTG + PB
8.2 ± 1.0b
12.0 ± 1.3
107.2 ± 10.8
106.8 ± 11.4
6.1 ± 0.8b
9.5 ± 1.1
56.1 ± 8.0b
80.8 ± 9.0
5.0 ± 0.7
6.9 ± 0.9
58.1 ± 7.1
54.7 ± 6.7
LTG + PHT
7.1 ± 0.8
8.5 ± 0.9
65.4 ± 4.4
70.9 ± 4.8
6.8 ± 0.7
7.1 ± 0.9
46.1 ± 4.3
56.9 ± 4.7
5.2 ± 0.7
5.8 ± 0.8
38.7 ± 4.2
42.8 ± 4.5
LTG + CBZ
10.9 ± 1.2
9.1 ± 0.9
38.5 ± 4.5
47.2 ± 4.2
11.5 ± 1.0b
7.6 ± 0.9
4.2 ± 4.6
41.0 ± 4.3
9.4 ± 1.0b
6.0 ± 0.8
30.5 ± 4.4
34.9 ± 4.3
LTG + TPM
21.9 ± 1.3
24.8 ± 1.5
528.5 ± 28.2
448.9 ± 30.6
12.9 ± 1.0a
18.2 ± 1.2
427.6 ± 22.4c
308.8 ± 21.8
9.4 ± 0.8
11.6 ± 1.0
192.8 ± 15.0
168.8 ± 13.1
Isobolographic analysis of neurotoxic effects evoked by lamotrigine combined with antiepileptic drugs in the chimney test in mice
TD50 values for all AEDs administered alone were determined in the chimney test in mice (Table 1).Applying the same isobolographic method for determining TD50 mix values and calculating TD50 add values for the AED combinations studied, it was shown that the interactions between LTG and VPA were either synergistic (1:3) or antagonistic (1:1; Fig. 1B). The supraadditive (synergistic) interaction was observed between LTG and PB only for the fixed ratio of 1:1 (Fig. 2B), whereas the combination of LTG with TPM (1:1) was evidently subadditive (antagonistic; Fig. 3B). The remaining fixed-ratio combinations between LTG and VPA (3:1) or LTG and PB (3:1) showed a pure additivity. However, combinations of LTG and TPM (either 1:3 or 3:1), although additive, showed tendency toward subadditivity (Fig. 3B). Conversely, the isobolographic analysis indicated a tendency toward supraadditive (synergistic) interactions for all fixed ratios of LTG with CBZ (Fig. 4B). With isobolography, combinations of LTG with PHT, for all fixed ratios tested in the chimney test, were additive (Table 2).
Dark-avoidance acquisition and retention testing
The control (vehicle-treated) animals did not enter the dark box within the observation cutoff time (180 s). Neither LTG alone nor coadministered with all studied AEDs, at the fixed-ratio 1:1 combination, affected long-term memory in the animals (results not shown).
Influence of lamotrigine on the free plasma concentrations of antiepileptic drugs and antiepileptic drugs on the free plasma level of lamotrigine
All experiments were conducted for the dose ratio of 1:1 (LTG to an AED). LTG at the dose of 1.4 mg/kg significantly reduced the free plasma level of PB (4.7 mg/kg) from 6.71 ± 0.6 μg/ml to 5.92 ± 0.5 μg/ml of plasma (Table 3). LTG at the doses ranging between 1.8 and 3.4 mg/kg did not affect the free plasma levels of the remaining AEDs (Table 3). In contrast, none of the AEDs studied affected the free plasma concentration of LTG evaluated with high-performance liquid chromatography (HPLC; Table 4).
Table 3. Influence of LTG on the free plasma levels of older AEDs and TPM
Plasma level (μg/ml)
Values are plasma concentrations (in μg/ml) of eight determinations in mice and expressed as means ± SD (standard deviation). Student's t test was used for statistical evaluation of plasma levels of AEDs studied.
Table 4. Influence of first-generation AEDs and TPM on the free plasma levels of LTG
Plasma level (μg/ml)
Data presented as plasma concentrations of LTG (in μg/ml) of eight determinations in mice, and expressed as means ± SD (standard deviation). Student's t test was used to perform the statistical analysis of data.
Results obtained from statistical calculations and isobolographic parameters
Both TD50 and ED50 values for all AEDs administered alone allowed the calculation of protective indices of the studied AEDs (Table 1). Moreover, for every combination of LTG with AEDs (at the fixed ratios of 1:3, 1:1, and 3:1, evaluated in the chimney and MES tests), a parameter characterizing the combinations in isobolographic analysis (i.e., BI) was determined (Table 2). Considering BI values for all AED combinations, LTG coadministered with TPM at the fixed-dose ratio of 1:1 provided the most advantageous combination, whereas BI values for LTG combined with CBZ at the fixed ratios of 1:3, 1:1, and 3:1 were 0.68, 0.55, and 0.56 respectively, pointing to very unfavorable interactions. Moreover, LTG combined with PB at the dose ratios of 1:3 and 3:1 offers some favorable effects because of increased anticonvulsant efficacy and simultaneously reduced neurotoxicity (BI for both combinations was 1.47). However, the combination of the drugs at the fixed ratio of 1:1 (due to synergy either in the MES or in the chimney tests) was disqualified as advantageous (BI was 1.08). Similarly, the combination of LTG with VPA at the fixed ratio of 1:3 was supraadditive either in the MES or chimney tests, with a BI of 1.18. Combinations may be clinically worth recommending, although they exert only a tendency toward synergy in the MES and antagonism in the chimney test, resulting finally in a high BI value. Such observations were evident for the combinations of LTG with TPM at fixed ratios of 1:3 and 3:1, with BI values of 1.33 and 1.41, respectively. All remaining combinations between LTG and the AEDs were classified as neutral on the basis of BI values (Table 5).
Table 5. Classification of AED combinations based on their interactions observed in preclinical studies
The anticonvulsant activity offered by AEDs in combinations is the first important factor considered by physicians in polytherapy. However, the side-effect profile of these combinations also is taken into account, when the adequate AED combinations for patients with intractable seizures are chosen.
Advantageous AED combinations:
• Synergy in terms of the anticonvulsant (therapeutic) effect
(TE) + antagonism as regards the adverse-effect profile (AEP),
the best combination
• Synergy in TE + additivity in AEP
• Additivity in TE + antagonism in AEP
Neutral AED combinations:
• Synergy in TE + synergy in AEP
• Additivity in TE + additivity in AEP
• Antagonism in TE + antagonism in AEP
Unfavorable AED combinations:
• Additivity in TE + synergy in AEP
• Antagonism in TE + additivity in AEP
• Antagonism in TE + synergy in AEP, the worst combination
The results shown from these experiments indicate a synergistic interaction between LTG and VPA or PB in the dose ratios of 1:3 and 1:1. In the ratio of 1:1, a synergy was found for combined treatment with LTG and TPM. In contrast, a distinct antagonism was noted for the combinations of LTG with CBZ coadministered in the ratios of 1:1 and 3:1. The remaining combinations of LTG with PHT (in all fixed-dose ratios tested) or with VPA, PB, or TPM at dose ratios not mentioned earlier, were additive. However, these statistical calculations are based only on data from the electroconvulsive test and do not consider adverse potentials of the drug combinations. Theoretically, a drug combination showing only additivity in a convulsive test but no or minimal adverse effects, also is relevant from a clinical point of view. To analyze both anticonvulsant activity and adverse potential of a drug combination, Luszczki et al. (26) proposed a new value, BI, which can simply be calculated by dividing the value of experimental protective index by the value of additive protective index, PImix/PIadd. When combining two AEDs with diverse PIs, a problem occurs in comparing the final PI of their combinations. Examining several fixed-drug dose ratios with different PIs, it is difficult to corroborate which combinations offer wide margins of safety. Questions arise about how to consider the combinations of which the PIs are lower than those of the AEDs applied alone or whether a PI for a combination at the fixed ratio of 1:3 is the same as for that of 3:1, etc. The BI allows much more insight into whether a combination of AEDs is attractive from both the anticonvulsant activity and adverse profiles. A very good example showing the predictive value of the BI is a combination of LTG with TPM. In terms of anticonvulsant activity and adverse effects, evaluated in the chimney test, only the combination at the dose ratio of 1:1 was synergistic or antagonistic. For the other dose ratios, the respective ED50 mix or TD50 mix values fell within the margin for additivity. Nevertheless, the respective BIs were always higher than the border value of 1.30. Conversely, a synergy was found between LTG and VPA (1:3), as regards anticonvulsant activity, but simultaneously also a synergy as regards motor performance.
According to Deckers et al. (8), a synergistic interaction is likely when the two drugs combined share different mechanisms of action. Additivity may be expected when the drugs combined share similar mechanisms. This assumption seems to be generally confirmed by the present study, because synergistic effects were observed between AEDs possessing various targets in the central nervous system (LTG and VPA, PB, or TPM; for review of their mechanisms of action see 8 and 34). Additivity was shown for two Na+ channel blockers, LTG and PHT. Conversely, one would expect additivity for LTG combined with CBZ, which was obviously not the case. From the analysis of the adverse activity in the chimney test, it may be postulated that there can be no theoretical considerations since synergy or antagonism was observed for AEDs displaying different profiles of mechanisms. This may be interpreted in terms of an obvious need for experimentally evaluated AED combinations for the subsequent rational polytherapy of epilepsy.
An important question emerges about a possible reason for the untoward interaction between LTG and CBZ. Recently Braga et al. (35) documented that LTG actually induces a substantial reduction in γ-aminobutyric acid subtype A (GABAA) receptor–mediated events in the basolateral amygdala. This effect has been evoked within the clinically relevant concentration of LTG of 10 μM. In contrast, an isolated study by Cunningham and Jones (36) indicates that LTG may enhance the frequency and amplitude of spontaneous inhibitory potentials by an elevation of GABA release in the rat entorhinal cortex. Some earlier articles suggested, however, that LTG actually reduces GABA release produced by veratradine in rat cortical slices (14,37). Certainly a clear-cut anticonvulsant effect of LTG is due to blockade of voltage-gated Na+ channels (11) and probably to inhibition of glutamate release (38), although according to Löscher (39), the latter effect occurs at supratherapeutic concentrations. A negative interaction may emerge when a drug sharing a similar profile of activity is combined with LTG. CBZ fulfils this criterion because this AED also has been shown to inhibit GABA release (37). Probably the antagonistic interaction is dependent on the impairment of GABA-mediated events by both LTG and CBZ, which is obviously not the case in other combinations.
Additionally, CBZ may behave as a nonselective adenosine-receptor antagonist (for review, see 40). Considering that adenosine-receptor antagonists (for instance, methylxanthines) are convulsant agents (41), this may be an additional reason for the antagonistic interaction between LTG and CBZ. PB to a certain degree also may be regarded as an adenosine A1-receptor antagonist (42) but this drug, in contrast to CBZ, considerably enhances GABAergic neurotransmission (8).
The antagonistic type of interaction between LTG and CBZ is not a unique one corroborated isobolographically for two Na+ channel blockers. Subadditive interactions, in terms of the anticonvulsant activity against MES in mice, were denoted either for oxcarbazepine (OXC) + PHT (26) or OXC + LTG combinations (unpublished data). In our previous study, it was observed that combination of CBZ with PHT exerted a strictly additive interaction [consistent with the study of Morris et al. (43)], whereas the combination of OXC with PHT was distinctly subadditive (antagonistic; 26). Conversely, the appearance of antagonistic interaction between LTG and OXC is quite similar to that of LTG and CBZ. As OXC is a derivative of CBZ, it is less likely that interactions between LTG and CBZ or LTG and OXC could be diverse. In contrast, interaction of LTG with PHT was purely additive in the present study. In light of these facts, it is likely that LTG influences Na+ channels as does PHT and, probably, differently from OXC or CBZ. Because a pharmacologic competition of drugs generally results in a pure additive interaction (8), no doubt remains that the combination of LTG with PHT is additive. Moreover, other interactions between CBZ and CBZ-10,11-epoxide (44) or CBZ and OXC (26) also were additive, which is in accordance with this hypothesis. Obviously, all diversities in the anticonvulsant activity of these Na+ channel blockers result from pharmacodynamic interactions between these AEDs.
Differences among interactions of Na+ channel blockers can be explained by different affinity of LTG and PHT or CBZ and OXC to the binding sites within the complex of Na+ channels, or by the existence of diverse Na+ channels with high affinity to CBZ or PHT. These presumptions might explain a molecular background and mechanisms responsible for this antagonistic phenomenon; however, some relations must be established between LTG and CBZ or OXC.
Another very intriguing hypothesis explains that LTG, owing to its molecular mechanisms of action, exerts an effect on a particular isoform of the brain Na+ channels or blocks “regionally selective” Na+ channels (13), and thus evokes its anticonvulsant activity. From a theoretic point of view, these hypotheses might, at least partly, explain the observed antagonistic interactions between Na+ channel blockers LTG and CBZ or OXC.
According to suggestions presented recently by Bourgeois (45), clinical evidence exists that antagonism may be found between AEDs applied in combinations. In a study examining the efficacy of vigabatrin (VGB) in children with infantile spasms, the efficacy of the drug was reduced in patients taking CBZ or VPA (46).
In clinical practice, it is very likely that an antagonistic effect exerted by two drugs in combinations (a pharmacodynamic antagonism) may be significantly masked or covered by a “false impression of synergy.” Generally in duotherapy, both AEDs are applied at the median therapeutically effective dose. With a pharmacodynamic antagonism between AEDs, both drugs are decreased in their anticonvulsant activity; however, the drugs in combination still offer a more pronounced antiepileptic effect than does each drug applied separately. Explanation of this phenomenon seems very simple. For instance, it is conceivable that one effective dose of the first AED added to one effective dose of the second AED should duplicate the final effect. However, because of the antagonism between AEDs, the final effect is reduced (for example) to 1.5 of the effective dose of the two-drug mixture of AEDs. It is evident that 1.5 ED still offers a significantly greater protection against seizures than does each AED administered separately in monotherapy; thus the antagonistic interactions may be clinically “synergistic.” In duotherapy, the final effect of such (antagonistic) combinations, observed by physicians, is advantageous because it is still greater than one effective dose of AED in monotherapy. When the therapeutic indices of the applied drugs were high, the adverse effects were never observed for this combined therapy, even if the antagonism occurred. In clinical practice, despite antagonism between AEDs, combinations of these drugs may offer complete seizure control. Hence, the combinations of LTG with CBZ or OXC with PHT may be clinically effective in patients with intractable seizures (4,9,10), although the drug load would be unnecessarily high.
It should be highlighted that direct extrapolation of results obtained from experimental studies on animals to clinical conditions is not possible. However, similar pharmacodynamic interactions occur between AEDs in animal models, which may help clinicians to choose the best AED combinations, based on rational polytherapy, providing the epilepsy patients with full seizure control at relatively low drug load (47,48).
It should be emphasized that these AED combinations were evaluated in one experimental model of epilepsy (MES test). With different tests [e.g., the lethargic (lh/lh) mice or PTZ-induced seizures], other results might have been generated for the anticonvulsant efficacy of these specific combinations.
It is obvious that the experimental data may only suggest the choice of AED combinations that might provide the efficacious treatment of patients with drug-resistant epilepsy. These experimental considerations, however, require clinical verification.
Some clinical reports indicate that the combined treatment with CBZ and LTG is considerably less efficacious than monotherapy with LTG in the management of resistant epilepsy (49,50), which seems to support our experimental data. However, the antagonism between LTG and CBZ has not been observed in many clinical trials. It seems likely that the absence of any signs of antagonistic interactions, in clinical practice, may generally result from the relatively high drug loads of AEDs in combinations. No doubt overtreatment, defined as an unnecessary and excessive drug load in the management of epilepsy, is a problem, especially in patients receiving polytherapy (45,51–55). Conversely, the clinically observed antagonism between VGB and CBZ or VPA, in children with infantile spasm, could result from an antagonism that is able to overcome the “masking additive effects” offered by excessive AED loads in combined therapy.
The combination of CBZ with LTG may also possess an undesired pharmacokinetic component, because Warner et al. (24) showed that adding LTG to CBZ results in an elevation of the primary CBZ metabolite [CBZ-10,11-epoxide) by 45%. This has led to the occurrence of signs of clinical toxicity [dizziness, nausea, diplopia) in almost 50% of patients. Similarly, Besag et al. (56) also reported on the existence of toxic effects (mainly diplopia and dizziness) of this combination in epilepsy patients, but they insist on the pharmacodynamic nature of this interaction because they did not observe any increases in plasma LTG or CBZ (including its metabolite) concentrations. Clinical trials also indicate that much better seizure control was evident when LTG was added to patients receiving VPA than with CBZ (57,58). Very similar results were obtained by Brodie and Yuen (59), who are of opinion that there is a clinical synergy between LTG and VPA.
However, a particular problem has appeared concerning the higher risk of congenital malformations in the offspring of mothers treated with LTG+VPA (60). Therefore, results obtained experimentally do not provide all major guidelines that must be considered for efficient and maximally safe antiepileptic therapy.
Pharmacokinetic events that might disturb the adequate classification of interactions of AED combinations in isobolography (61) were excluded. In spite of the reduction of the free plasma level of PB, after adding LTG, a pharmacodynamic character of this interaction persisted. The isobolographic supraadditivity for this LTG+PB combination testifies in favor of a pharmacodynamic character of the interaction, leaving no doubt that the observed interactions between LTG and AEDs tested are of a pharmacodynamic character.
It is important to note that for the first time, synergistic types of interactions between LTG and TPM or VPA, denoted isobolographically in animals, also are effective in the therapy of patients with refractory epilepsy (62–64). Considering thoroughly the separate mechanisms of action of each component in combination, it is evident that both drugs cooperate synergistically in the reduction of seizure activity.
Generally monotherapy starts with a first-line older AED. When monotherapy fails, the patients are switched to another (alternative) monotherapy with a newer AED. This is a main source of evidence of the effectiveness of two-drug combinations. It is clear that majority of data exist only for combining older with newer AEDs. With the advent of eight second-generation AEDs, introduced in the 1990s into the epilepsy armamentarium, a growing body of evidence is becoming available about the effectiveness of newer AEDs in combinations, especially for LTG and TPM, which seem to be the most efficient drugs in combination. Overall, this combination (LTG+TPM) fulfills all criteria (theoretic, experimental, and clinical assumptions of rational polytherapy) to be the most effective. In the MES test in mice, this combination exerted synergy; its adverse-effect profile in the chimney test was subadditive (antagonistic), finally giving a high BI. In clinical practice, this combination has some merits in protecting patients with refractory epilepsy (64). We hope that increasing clinical evidence will confirm or reject our presumptions, verifying the usefulness of the BI as a measure factor of the efficacy of AEDs in combinations from animal studies to human conditions (Table 5).
Another very promising combination is possible of LTG with the newer AED, gabapentin (GBP). In our previous study, it was ascertained that the combinations of LTG with GBP, at various fixed drug-dose ratios, were synergistic, providing a significant antiseizure protection against electroconvulsions in mice (32). In spite of only one clinical case reporting the effectiveness of these AEDs combined in a patient with intractable seizures (4), the theoretic presumptions and preclinical studies clearly indicate that LTG and GBP could be rationally combined to potentiate the antiseizure effects offered by these AEDs in patients inadequately controlled with monotherapy.
Very beneficial interactions between LTG and TPM or LTG and VPA in terms of both anticonvulsant activity and adverse effects must be accentuated. Conversely, a substantial antagonism as regards the combination of LTG with CBZ indicates that this combination of AEDs should be avoided. Some clinical data support this assumption (50).
Acknowledgment: We are grateful for the generous gift of valproate magnesium from ICN-Polfa S.A. (Rzeszów, Poland) and carbamazepine from Polfa (Starogard, Poland). This study was supported by grant 6P05F 026 20 from the State Committee for Scientific Research, Warsaw, Poland. Dr. Luszczki is the recipient of a Fellowship for Young Researchers from the Foundation for Polish Science