I'll be talking about the off-label use of valproate, clonazepam, benzodiazepines, ethosuximide, prednisone, adrenocorticotropic hormone (ACTH), and intravenous gamma globulin.

We've heard some talk tonight about status epilepticus and convulsive and nonconvulsive status epilepticus. I refer to what's been discussed as overt status epilepticus. This would be the status epilepticus syndromes that are associated with acute seizures or an acute encephalopathy.

I call some of these epileptiform encephalopathies. These would be specific epileptic syndromes that are associated with frequent epileptiform activity on the electroencephalogram (EEG). Frequently, we may see these during sleep, so we may have an EEG that becomes very sleep-activated. This would be called electrical status epilepticus of sleep (ESES). There's also an epileptic syndrome that's called ESES, which is important to differentiate from an EEG, which may be a very sleep-activated EEG. These can be referred to as epileptiform encephalopathies.

I use epileptiform encephalopathy to define something in which the primary clinical manifestations result from the epileptic activity, the epileptiform features on the EEG, and its resultant dysfunction, rather than an actual clinical seizure. Sometimes it can be difficult in differentiating these.

What we're focusing on are patients who may clinically not have many seizures, but have EEGs that are very epileptiform EEGs. Not all of these patients may have overt clinical seizures.

One thing that can be very important in these disorders is regression in either intellectual or cognitive ability. If you see regression in either intellectual or cognitive abilities, that has to raise the suspicion that we may be dealing with some of these very sleep-activated epileptiform encephalopathies.

These disorders are seen in children. The first is the Landau-Kleffner syndrome, which is a disorder characterized by a previously normal child in whom there's an acute loss of language function. These can be associated with seizures. They can be associated with behavioral problems. The key is someone who has had regression in their language function.

There's the epileptic syndrome that's called ESES. That's a disorder in which there's perhaps a cognitive or intellectual decline but not just a language decline. These patients tend to have an overall intellectual decline with regression but not specifically in their language function.

There are other disorders that can be similar to this. One is the pervasive developmental disorder, or the autistic patient who has regression. It's possible that some of these patients may have very sleep-activated EEGs, and the question becomes, what do you do with these patients?

The childhood disintegrative disorder can have a similar finding on the EEG. Patients who have a developmental language disorder, a congenital aphasia â€' some of these patients may not have developed normally â€' they may have never had a seizure, yet when you've looked at their EEGs, you can see very sleep-activated EEGs.

The last group I want to talk about, and raise the question of how do you treat them, is the group of patients with transient cognitive impairment. We've typically looked at EEGs that have long bursts or short bursts. If it's not associated with a clinical symptom, or a clinical manifestation, we've said that this isn't a seizure or epilepsy. Yet if some detailed testing can be done on these patients, such as computerized, verbal, or visual memory testing, you may see, during some of these bursts, that there are clinical manifestations associated with it. You get into the question of how aggressively do you treat these patients?

The syndrome of epilepsy with continuous spikes and waves during slow sleep is ESES. You have an 85% to 100% spike index during non-rapid eye movement (REM) sleep. This means that when you're in non-REM sleep, you have to have 85% spikes. I don't think you absolutely have to have 85% of your slow wave sleep as spikes in order to say that the person has regression or qualifies for ESES. There are some criteria that call for a Type 1 or Type 2 ESES depending on the amount of epileptiform activity on the EEG. Although some of the cases of ESES may not have spikes during REM, the spikes may persist during REM sleep.

In the waking EEG in these patients, you tend to see diffuse, 2- to 3-Hz spike-and-wave discharges occurring in bursts. There can be focal spikes so they can have spikes that occur frontally, centrotemporally, or even posteriorly. These patients may have epilepsy or associated seizures, but they have a neurobehavioral disturbance with intellectual regression. The hallmark of these disorders may be some type of regression.

When we talk about Landau-Kleffner syndrome, or ESES, we tend to divide things into how they're affected etiologically, and of course, idiopathic vs symptomatic. The group that has had the continuous spike waves of slow sleep has been divided into the idiopathic and symptomatic, with the idea that the idiopathic cases have a normal psychomotor development to begin with. This is followed by intelligence quotient (IQ) loss. There may be language regression in these patients, but it's not a specific or more isolated language regression like you might see in Landau-Kleffner syndrome. That's not to say that Landau-Kleffner syndrome can't have some behavioral disorders associated with it. It's actually very common when children have regressed to have associated language regression.

There can be impaired temporospatial orientation, and they may have hyperactivity with aggression in psychotic states. The difference would be patients who are abnormal to begin with neurologically. They can have an abnormal neurologic examination. They've had abnormal psychomotor development. And this has been followed by the same findings: IQ loss, language regression, or marked aggression. Coming back to etiology and prognosis, the group that is idiopathic may do better in the long term than the group that is symptomatic.

For the treatment of these disorders, particularly Landau-Kleffner syndrome, these patients should be treated with intensive speech therapy and a language-based program. How do you treat them after this? What's difficult with some of these disorders is that they're rare, and it's hard to put together studies to study these disorders. For antiepileptic drugs for Landau-Kleffner syndrome, a first choice would be a spike suppresser, something like valproate. Some of the anticonvulsants aggravate some of the disorders. Carbamazepine can aggravate Landau-Kleffner syndrome.

I might use something like valproate as a first-line therapy for these patients. If that doesn't work, you may try corticosteroids. The question would be do you use prednisone or do you use ACTH? Patients have been treated with intravenous immune gamma globulin, even people who have had some underlying immune disorder. In some selected cases, epilepsy surgery may be appropriate, depending on a symptomatic case with a certain kind of lesion. It could be a lesionectomy, if you had a patient with a focal lesion; or you could do a multiple subpial transection, which is a procedure for removing an epileptic focus within eloquent cortex, defined as cortex that you wouldn't want to resect because you'll lose some important function, like motor function or language function.

For ESES, it's basically the same list of treatments, because Landau-Kleffner syndrome is associated with sleep-activated EEGs, and some centers require ESES to make a diagnosis of Landau-Kleffner syndrome. Personally, I don't think that you have to have ESES, or continuous spike waves of sleep, in patients with Landau-Kleffner syndrome, but you certainly need to have a sleep-activated EEG.

The standard anticonvulsants that have been used most, or have been the best, for ESES have been spike suppressers such as valproate, clonazepam, or ethosuximide. I haven't included some of the newer anticonvulsants, but medications like lamotrigine or topiramate have had some response in these conditions.

The same question in corticosteroids. Do you use prednisone or do you use ACTH as a first-line therapy? Because of the major behavioral or attentional difficulties that a lot of these patients have, psychopharmacological agents can be particularly helpful, such as something to enhance their attention or take care of their impulse control. Some of this could be ameliorated or treated by controlling the epileptiform activity on the EEG itself. Intravenous immunoglobulin has been used in some of these patients and, also, epilepsy surgery with either multiple subpial transection or a lesionectomy.

This is a slide of ESES, or the other term that has been used is continuous spike waves of slow sleep. You see a continuous discharge. It's generalized. It's occurring at a frequency of about 2 to 3 Hz. And, of course, it's generalized and present during sleep.

This next EEG is not the same EEG, but it's similar. The difference is the sensitivities. The first EEG was one where the sensitivity was not as high. This was done at a bar of 700 mcV. You can see very high electrical activity, and you do that to prevent the blocking that we saw on the past EEG. So this would be an EEG during sleep of ESES.

How do we treat these patients? For Landau-Kleffner syndrome, this is an initial EEG. You can see the awake EEG. The person has left temporal occipital spikes and sharp waves. When they go to sleep, they have ESES, or continuous slow wave sleep. So this was both the waking and the sleep EEG.

Initially we can treat some of these patients with standard anticonvulsants, but our data have shown that these patients don't generally do well with standard anticonvulsants as the first-line therapy. In this patient, we had used prednisone, and you can see this is the same sleep-activated EEG. One month after high-dose prednisone therapy -- we've been using oral prednisone at 2 mg/kg as an initial dose â€' this person has a resolution of the ESES and still has some slowing in the temporal regions bilaterally, greater on the left. But 1 month later, the person does have control of the ESES, and in this patient, we did begin to see some improvement in his language and his behavior.

This is his EEG after 6 months of high-dose prednisone therapy done in a tapering schedule. You can see that at this point, the EEG is asleep and there are vertex waves. This was a normal sleep EEG with prednisone.

This is the same slide we saw earlier of ESES. After prednisone, we see a resolution of the ESES. But this person continued to have, at least in the follow-up EEG, some residual spikes in the frontal region.

This is another patient who had the syndrome of ESES. This girl had infrequent seizures, but she had a behavioral and a cognitive decline where she couldn't remember anything, and she was declining in her schoolwork. She was put on carbamazepine first because of focal seizures. Her EEG worsened. She then was placed on valproate. Her EEG didn't improve. Then we treated her with prednisone, and this is her EEG pretreatment with prednisone. When she's asleep, she has ESES.

One month later, when there had been some behavioral and cognitive improvement, the EEG is improved, but she still has residual spikes. So, after a month, we didn't see a complete resolution of her EEG. She still had residual spikes. The spikes look like they're occurring at about the same frequency that they were occurring previously, but the amplitude of the spikes is markedly decreased.

She went on, during the taper of the high-dose prednisone, to have another period of regression. Then we put her back on high-dose prednisone and tapered her more slowly, the second time successfully.

This is work that was presented by Albaradie. We had 12 patients with follow-up of at least 18 months. There were 6 males and 6 females. Seven were symptomatic; 5 were idiopathic. Five had Landau-Kleffner syndrome with primarily a regression in their language, whereas 7 had the syndrome of ESES.

These patients all were treated with a prior anticonvulsant. When we looked at those who responded after the addition of a second antiepileptic drug, which was typically valproate, only 1 out of the 12 had a resolution of the ESES with the addition of valproate.

Ultimately 6 were treated with prednisone, and 2 were treated with a high-dose diazepam (Valium) protocol. Of the 6 treated with prednisone, the ESES disappeared in 5 but returned in 4 out of the 5. The patients needed, therefore, a second period of treatment with prednisone. But prednisone, at least in our data, looks like it's effective in controlling the EEG abnormalities. We had 1 patient in whom a lesionectomy was done, and we had 1 patient in whom a multiple subpial transection was done in association with an epileptic resection.

We've also looked at a different protocol for the treatment of continuous spikes and waves during slow sleep, a high-dose diazepam (Valium) protocol. This was the treatment of the ESES syndrome with high-dose diazepam. We were using diazepam at an initial dose of 1 mg/kg, and the first dose is typically given rectally. After that, a dose was given orally at night, from 0.5 to 0.75 mg/kg, and then EEGs were done in follow-up. We had 14 patients with follow-up of at least 3 months. Three of these were patients with Landau-Kleffner syndrome, and 11 were patients who had ESES.

For the Landau-Kleffner syndrome group, 3 out of 3 were improved with the high-dose diazepam (Valium) protocol. In those who had ESES, there was an initial clinical response, an EEG response in 8. After 3 weeks, 50% had EEG improvement, but only 25% had had an improvement in the seizures.

The side effects were surprisingly low. You would think if you give somebody a high-dose diazepam (Valium) protocol of 1 mg/kg that they'd have a lot of side effects, but we saw side effects in only 4 out of 14. These were mostly behavioral problems and lethargy. When we saw these, we were reducing the dose somewhat, but there were no patients we had to take off of the high-dose diazepam (Valium) protocol because they couldn't tolerate it.

This work came out of a paper by Dr. De Negri in which they looked at 43 children, aged 5 months to 14 years, with ESES and treated them with rectal diazepam at 1 mg/kg. This protocol is typically done with the patient under EEG monitoring. In the patients who were responders, they were then given 0.5 to 0.75 mg/kg/d orally over 3 to 4 weeks. He defined a positive response as either complete remission or a greater than 50% reduction in the epileptiform activity.

In this group, some of the patients did need to be treated with a subsequent course, and they thought that they could follow them clinically. When the patients regressed clinically, that was time to treat them again. In the early data that we have, it's looking like the EEGs may regress prior to a clinical regression, so we may need to do more frequent EEGs in following up these patients.

I want to include some things about surgery for Landau-Kleffner syndrome or ESES. I mentioned lesionectomy earlier. Multiple subpial transection can be done for this. We've done vagal nerve stimulation in 3 of these patients with ESES, and we haven't seen any significant change in the epileptiform activity on the EEG.

This shows the multiple subpial transection that was invented to disrupt, but not disconnect, an epileptic focus. You disconnect the horizontal cortical fibers from one another but leave the vertical columns intact so you're not denervating the function. This way you can perform a procedure in eloquent cortex and not lose function.

This is one of our interesting patients. This is a 7-year-old boy who started out with left-sided hand seizures and was started on various anticonvulsants. Over the course of about a year, he had an increasing frequency of his seizures, consisting of staring spells and eye blinking, and was refractive to anticonvulsant therapy.

He had magnetic resonance image (MRI) initially, which was supposedly normal. But when we went back and looked at this MRI, there was this subthalamic mass. You can see it on the right side, but not on the left. His EEG showed ESES. The sensitivity on this has been markedly decreased, but he had ESES with a right-sided predominance.

If you look at this again, you can see this lesion. After he was refractory, we went in and did a lesionectomy. Because of its location, the surgeons were very reluctant to take out the whole lesion because it went up into his thalamus. So we took out the portions of this lesion which were easily removed, and he went seizure-free for about 6 months. The pathology was a glioneuronal hamartoma. He went seizure-free, and within about 3 months had a normal EEG. But 6 months later, he started having seizures again and then went on and was more refractory. We ultimately did a complete lesionectomy, and he's now been seizure-free for about 3 years. That was a case of someone who had lesional ESES.

This case is a 4 year-old boy who presented with febrile seizures. He had 2 febrile seizures, was otherwise fine, but then within a couple of months began to lose language. He lost his receptive language first and then his expressive language and went on to develop auditory agnosia. We thought he had Landau-Kleffner syndrome.

When we looked at his first EEG, it had ESES. He had had a waking EEG that had bitemporal spikes, but he did not have a sleep EEG. So the ESES was not picked up immediately. Whenever we're evaluating these kind of patients with language regression, you really need to obtain a sleep EEG to assess for some of these disorders.

He was treated first with carbamazepine, then clonazepam, and he got worse. He kept getting worse. We put him on valproate and he started having staring spells and blinking. Those episodes markedly increased in frequency when we put him on valproate, which doesn't typically happen.

He was started then on high-dose prednisone, the prednisone protocol for Landau-Kleffner syndrome, and he didn't respond to that. We decided to give him the high-dose diazepam (Valium). This was his EEG immediately before diazepam (Valium), and you can see the pattern that we would consider ESES. He was given 1 mg/kg rectally, and this was his EEG immediately after the diazepam. He had a marked decrease in the amount of epileptiform activity on his EEG.

This is an EEG further in the same night. Again, you can see that there's been a marked improvement in his EEG. We kept him on 3 weeks of high-dose diazepam (Valium), and this was his sleep EEG after 3 weeks. You can see, with sleep, there are vertex waves and sleep spindles. He had a normal EEG after 3 weeks of high-dose diazepam (Valium) therapy.

Clinically, he was doing very well. The family thought, about a year later, that he seemed to sometimes have word-finding difficulty, and I repeated an ambulatory EEG on him. He, again, had a sleep-activated EEG. He didn't have ESES the first time, but he had bilateral central parietal spikes. By the time we monitored him, he had ESES, and we gave him high-dose diazepam (Valium) again. He responded, and, again, has a normal EEG. The high-dose diazepam (Valium) worked, but he regressed with the therapy.

Case 3 is a girl with language regression who never had any seizures. This was her first EEG done in the waking state in which you can see these temporal discharges. When she went to sleep, the temporal discharges increased in frequency.

She was started on valproate and actually got worse on valproate. She was then started on zonisamide and had a reaction to zonisamide.

We brought her into the hospital for monitoring. This was her waking EEG with some temporal spikes. When she went to sleep, she had the pattern of ESES. This is a portion from the same EEG before any therapy. So you can see that these patients may not have 85% of their EEGs as spikes, but they still have the same behavioral or cognitive regression.

We gave her a dose of high-dose diazepam (Valium) rectally, and you see her EEG after 24 hours. This is a sleep EEG in which we just see this beta activity.

In some of these disorders that have marked epileptiform activity on the EEG, we may want to be thinking about how aggressive we should be in treating these patients.