CBZ 2jjj


50 mV 200 ms

Figure 16.8 Cellular action of phenytoin and carbamazepine. Each column shows the response of a spinal cord neuron in culture to four increasing directly applied current pulses (amplitude in nA given at start of each sweep. Under control conditions (CONT)) the progressive depolarisations (bottom to top of each column of traces) induce increasing sustained discharges, whereas in the presence of phenytoin (PTZ) and carbamazepine (CBZ) firing cannot be maintained although the initial action potential remains. (Reproduced from MacDonald and McLean 1986.) These drugs are thought to bind to Na+ channels after they have been active (opened) and maintain them in the inactivated state. Thus they do not affect the initial response but stop neurons from maintaining the abnormal sustained discharge that would be characteristic of epileptic activity. Resting membrane potentials (Em) are shown at the bottom of each column and amplitude (mV) and time (ms) at the bottom right with phenobarbitone, can precipitate seizures. Although still used in refractory myoclonic epilepsy, when its depressant effect on the spinal cord may be significant, clonazepam, like phenobarbitone, is rarely used now, but the more recently introduced 1:5 benzodiazepine clobazam is quite often used as an adjunct (not in the United States). While there is some belief and evidence that clonazepam and clobazam are more effective than other benzodiazepines as anticonvulsants nothing is known specifically about their modes of action that supports this view. The reported inhibitory effects of B & Bs on a calcium-sensitive NT release in synaptosomes is difficult to evaluate in terms of their in vivo anticonvulsant activity.

Valproic acid (sodium valproate)

Introduced initially for absence seizures, this drug is now known to be effective in and used to treat tonic-clonic seizures and most types of epilepsy. It was found to inhibit GABA transaminase and so elevate GABA concentrations and inhibition. This is achieved, however, over a slower time-course than its anti-seizure effect, especially experimentally, which is now thought to be due to its phenytoin-like, use-dependent block of sodium channels. Since, unlike phenytoin, the full effect of valproate takes some weeks to develop, its slower effect on GABA metabolism and activity should not be ignored.


Most of these have been used mainly as add-on therapy although some are now being used alone.


One unwanted side-effect of phenytoin is its anti-folate activity. A programme of synthetic chemistry to manipulate the structure of the anti-folate compound pyri-methium to try to replace that property with anticonvulsant activity resulted in the synthesis of lamotrigine. It proved to be an effective AED in partial and generalised epilepsy but experience has found it also to be of value in absence seizures.

Experimentally it was shown to reduce the release of glutamate and to a lesser extent GABA, induced in small brain slices by veratridine, a sodium ion channel opener. It now appears that its primary effect is prolonging the inactivation of sodium channels in a use-dependent manner much like phenytoin, although in a recent study of intra-cellularly recorded activity of striatal neurons in the rat corticostriatal slice preparation some differences emerged. While both drugs reduced experimentally induced repetitive firing, phenytoin was more effective against those induced by direct current activation of the neurons and also inhibited the EPSPs induced by the direct application of glutamate. By contrast, lamotrigine had little effect on the glutamate response but was more active against those induced by corticostriatal tract stimulation, suggesting that part of lamotrigine's action may still reside presynaptically in reducing glutamate release (Calabresi et al. 1999).

Vigabatrin (y vinyl GABA)

This drug is chemically related to GABA, is an irreversible inhibiter of GABA transaminase and appears to produce its antiepileptic effect through that mechanism. Not only does it increase brain GABA levels in animals it also elevates them up to threefold in human CSF and in the occipital cortex of normal and epileptic patients as shown by nuclear magnetic resonance spectroscopy. An interesting decrease in glutamate may be secondary to the rise in GABA. It is effective in partial and secondary generalised epilepsy, but since its mode of action requires the regeneration of new enzyme (GABA-t) its effect far outlasts its plasma life. A worrying intramyelinic oedema in rat nerves has fortunately not been seen in humans or primates.


Drugs that block the neuronal and in particular the glial uptake of GABA, like diamino-butyric acid and nipecotic acid respectively, proved effective anticonvulsants experimentally but had to be administered directly into the ventricals (intra-cerebroventrically). Attaching nipecotic acid to a lipophilic component to increase brain penetration resulted in tiagabine. Surprisingly, it appears to act preferentially on the GABA transporter GAT1 which, although found on astrocytes, is more associated with nerve terminals. Microdialysis in rats shows it increases extracellular GABA and prolongs the post-excitatory hyperpolarisation of neurons. It has proved effective in partial and secondary generalised epilepsy but prolonged post- and possibly presynaptic actions of the increased GABA could present problems.


This drug, which is a cyclohexone analogue of GABA, was synthesised in the hope that it would be an agonist for GABA receptors which could cross the blood-brain barrier. Its efficacy in drug-resistant partial and secondary generalised epilepsy means that it certainly must enter the brain but it does not bind to GABA receptors. Despite this, it appears to increase GABA brain levels in epileptic patients and weak potentiation of GAD and inhibition of GABA-t have been described. It does not appear to affect sodium or calcium channels even though experimentally chronic dosing blocks repetitive neuronal firing. Specific binding sites have been shown for it on neuronal membranes which appear to be a leucine transporter, but their significance is not clear.


The last few years has seen an explosion in AEDs. Some of those mentioned above may fall by the wayside and others appear. At the time of writing, we could include felbamate, zonisamide oxcarbazepine and topiramate. They all appear to have a phenytoin-like action on sodium channels, although topiramate appears to also potentiate the action of GABA on GABAA receptors like the benzodiazepines but through a different site.


It will be apparent that all the possible mechanisms of action for anticonvulsant drugs outlined above (Fig. 16.6) have not been realised by those drugs currently available. The efficacy of glutamate NMDA antagonists is still restricted to experimental studies. No clinically useful drug has been developed and its synthesis will depend not only on finding a compound capable of entering the brain but also on the realisation of the hope that focal NMDA receptors may prove to be different from others. It may then be possible to target them specifically and avoid widespread depression. Lamotrigine does reduce the release of glutamate but this may be secondary to the blockade of sodium channels.

No directly acting GABAA receptor agonists have been found and it is likely that they would be too depressant (widespread in action) unless focal GABA, like NMDA, receptors have undergone some changes to become specifically targetable. Drugs that decrease the destruction of GABA such as GABA-t inhibitors (vigabratrim) and uptake blockers (tiagabine) have, however, been developed.

Despite all these approaches, drugs acting directly on neuronal ions channels are still the most effective AEDS.

Whether one drug with one mechanism of action will ever be adequate in the therapy of epilepsy is uncertain. Even drugs which apparently have a similar mechanism of action on sodium channels, such as phenytoin, carbamazepine, valproic acid and lamotrigine have different uses as only the latter two are effective in absence seizures. This could reflect some action additional to that on sodium channels (e.g. GABA-t inhibition for valproate) or an effect on a particular type of sodium channel that is different by virtue of some change in its a subunits. In fact the additional clinical effect of some new AEDs (e.g. vigabatrin and tiagabine) in patients not properly controlled by old AEDs like phenytoin could indicate the need for increased GABA function as well as sodium channel block for proper seizure control. The obvious complexity of NT and ion channel interactions in the control of neuronal function may well mean that the proper control of seizures may require the appropriate manipulation of more than one NT and one neuronal function.

Newer AEDs do have some advantages in that they tend to have fewer effects on the metabolism of each other or other drugs. By contrast, phenobarbitone is one of the most potent inducers of the microsomal enzyme system (cytochrone P450) responsible for the metabolism of drugs. Phenytoin and carbamazepine have a similar but less marked effect while valproate inhibits the system.

One thing is certain. All the new AEDs are much more expensive than the older ones and one might therefore question the justification of their use. The reason is that the older ones have limited efficacy and not-inconsiderable toxicity. Indeed even with polytherapy the seizures are not always adequately controlled. So are there other approaches?


If there is a clear established focus then maybe the best treatment is to remove it. This is, of course, both difficult and expensive but its use is expanding with about 500 operations per year in the UK. It is only considered in cases of partial (not general) epilepsy when conventional drug therapy has failed and a clear focus can be established. The advent of sophisticated assessments, such as MIR, long-term EEG telemetry, in-depth electrode recording and PET studies of blood flow and diazepan binding has now made this possible. Most commonly part of the anterior temperal lobe is removed, 70% of patients become seizure-free and neurological (mainly visual) and psychiatric problems are surprisingly few (5-10%).


This is not really a treatment but there is a view that glial cells can protect against seizures since the enzyme systems they possess (e.g. Na-K+ATPase and carbonic anhydrase) facilitate the regulation of ion movements and reduce the spread of seizures. Certainly ageing, a fatty diet, and phenytoin itself increase glial cell count while decreasing seizure susceptibility. In fact inhibition of carbonic anhydrase and the production of bicarbonate was one of the first treatments for epilepsy and a recent discovery that under certain circumstances intracellular bicarbonate can depolarise neurons has created a fresh interest in it.

Continue reading here: References

Was this article helpful?

0 0