Monday, 19 January 2015

Modified Use of Anti-Epileptic Drugs (AEDs) at Low Doses in Autism

As readers will be aware, many people with more severe autism are also affected by epilepsy.  Siblings of those with autism also seem to be at greater risk of epilepsy.

There are frequent comments that once starting on AEDs (Anti-Epileptic Drugs) aspects of autism also seem to improve.  This should not be surprising given the suggested action of these drugs and the overlapping causes of epilepsy and autism.

Today’s post is prompted by the observation that in very low, apparently sub-therapeutic, doses some AEDs seem to improve autism in some cases.  This is relevant because the usual high doses of these drugs are associated with some side effects and indeed a small number can be habit forming.

What is epilepsy?

The cause of most cases of epilepsy is unknown.

Genetics is believed to be involved in the majority of cases, either directly or indirectly. Some epilepsies are due to a single gene defect (1–2%); most are due to the interaction of multiple genes and environmental factors.  Each of the single gene defects is rare, with more than 200 in all described.  Most genes involved affect ion channels, either directly or indirectly. These include genes for ion channels themselves, enzymes, GABA, and G protein-coupled receptors.

Much of the above applies equally to autism, including the genetic dysfunctions associated with GABA.  The ion channel dysfunctions in epilepsy are thought to be mainly sodium channels, like Nav1.1.  We previously came across this channel when looking at Dravet Syndrome.

Dravet Syndrome

Dravet Syndrome is rare form of epilepsy, but is highly comorbid with autism.  It is cause by dysfunctions of the SCN1A gene, which encodes the sodium ion channel Nav1.1.  There is a mouse model of this condition, used in autism research.  Dravet Syndrome is known to cause a down-regulation of GABA (the neurotransmitter) signaling.  We saw how tiny doses of Clonazepam corrected this dysfunction in mice.

Known ASD-associated mutations occur in the genes CACNA1C, CACNA1F, CACNA1G, and CACNA1H, which encode the L-type calcium channels Cav1.2 and Cav1.4 and the T-type calcium channels Cav3.1 and Cav3.2, respectively; the sodium channel genes SCN1A and SCN2A, which encode the channels Nav1.1 and Nav1.2, respectively; and the potassium channel genes KCNMA1 and KCNJ10, which encode the channels BKCa and Kir4.1, respectively.

Dr Catterall, the researcher, then went on to test low dose clonazepam in a different mouse of autism model and found it equally effective.  It also appears to work in some human forms of autism.

Sodium Valproate

Valproate is a long established epilepsy drug that has also been used widely as a mood stabilizer and particularly to treat Bipolar Disorder.

One side effect can be hair loss.  Hair loss/growth and also hair greying are frequently connected with drugs and genes linked to autism (BCL-2, biotin, TRH etc).

One regular reader of this blog has pointed out that a tiny dose of Valproate, when combined with Bumetanide, appeared to have a significant and positive effect.  We know that bumetanide works via NKCC1 and the GABAA receptor to make GABA more inhibitory.

Many modes of action are proposed for Valproate, but the most mentioned one is that it increases GABA “turnover”; so it would make sense that having shifted the balance from excitatory to inhibitory, a stimulation to increase GABA signaling might be beneficial.

What is odd is that this is happening at a dose 20 times less than used in epilepsy, bipolar or mood disorders.

The use of Clonazepam, discovered by Dr Catterall, is also at a dose 20 to 50 times less than the typical dose.

Clonazepam and Valproate are both AEDs.  There are not so many of these drugs and while using them at high doses, without dire need, might be highly questionable, their potential effectiveness at tiny doses is very interesting.

Clonazepam is a Benzodiazepine in the table below.

The above table is from the following paper:-

Low Dose Clonazepam

Low dose Clonazepam was shown to be effective by its action of modulating the GABAA receptor to make it more inhibitory.  There are different types of GABAA receptor and the low dose effect was sub-unit specific.  Other benzodiazepine drugs were found to have the opposite effect.

The mouse research showed that the effect only appeared with a narrow range of low dosages.

Low Dose Valproate

Valproate is known to affect sodium channels like Nav1.1, but also some calcium channels.

For an insight into some known potential effects of Valproate, here is a paper from the US National Institute of Mental Health:-

In the paper it highlights the less well known effects of Valproate:-

inhibits HDACs
Modulates Neurotrophic and Angiogenic Factors (BDNF, GDNF, VEGF)
PI3K/Akt Pathway
Wnt/β-Catenin Pathway
MEK/ERK Pathway
Oxidative Stress Pathways
Enhanced Neuroprotection
Enhancing the Homing and Migratory Capacity of Stem Cells

Here is a list of the suggested new applications of Valproate, many highly appropriate to many types of autism:-

*       A. Stroke
*       c. Anti-inflammation
*       d. Angiogenesis
*       e. Neurogenesis
*       b. Anti-inflammation
*       c. BBB protection
*       d. Angiogenesis
*       e. Neurogenesis
*       B. TBI

Having read that paper I am now not surprised that a tiny dose of valproate can have a positive behavioral effect in autism.  What would be interesting to know is how the effects and dominant modes of action vary with dosage.  I presume the dosage has been optimized to control/prevent seizures.

Valproate is a cheap drug and is available as a liquid, so accurate low dosing is possible.  It has been shown to be neuro-protective, even shown promise as a treatment for traumatic brain injury.

While not written about autism, some of you may find the following collection of research interesting:-

It does talk about the wider potential use of Valproate, but not at tiny doses.


Interestingly, an orphan drug was developed in the European Union to treat Dravet Syndrome.  It is included on the list of AEDs above.

Even though that drug, Stiripentol, is not approved by the FDA, most sufferers in the US are able to acquire it under the FDA’s Personal Importation Policy(PIP).

So it is indeed possible to acquire drugs prior to approval in your home country.

Hopefully, once Bumetanide is approved for autism in Europe, similarly people will be able to access it easily in the US.

I wonder if anybody with Dravet Syndrome has tried low dose Clonazepam.  In theory it should be helpful.


  1. Vinpocetine is a anti-seizure drug that has an interesting cognitive effect on my son (which has no seizure history).
    João Santos

    1. An interesting article that states "Vinpocetine improves neuronal plasticity and reduces the release of inflammatory cytokines and chemokines from endothelial cells, vascular smooth muscle cells, macrophages, and microglia, by inhibiting the inhibitor of the NF-κB pathway".
      "Anti-Inflammatory Effects of Vinpocetine in Atherosclerosis and Ischemic Stroke: A Review of the Literature" at
      João Santos

    2. Thanks João, this is very interesting. Most AEDs seem to be sodium channel blockers, but some are entirely different. I am looking at the NF-κB at the moment, there are 700+ drugs that have been found to affect it, so it would be hard to know where to start. "Stumbled-open" is likely the best method.

      What dose of Vinpocetine are you using and what is the effect?

      Piracetam is another AED that does not affect Na channels. It was trialed in schizophrenia and improved things, but only in terms of cognitive performance. I tried it and found it to be effective in a similar way (he started making simple jokes).

      The good thing for many people is that while Vinpocetine is a drug in Europe, it is a supplement in the US. As you might expect, most readers of this blog are from North America.

      Any other drugs that you have found effective?

    3. Here is an interesting article published recently:
      The treatment of autism with low-dose phenytoin: a case report
      It talks about low dose Phenytoin

    4. Very interesting and the article also refers to low dose Valproate use.

      Significantly, this beneficial effect of sodium valproate appeared to have a narrow therapeutic window, with the optimal range between 50 and 200mg daily. A complete loss of efficacy frequently occurred above a dose of 400mg.

    5. Note on that paper the section at the bottom:-

      Competing interests

      The author owns shares in AutRes. This company owns patents and intellectual property that include the treatment of ASD and ADHD with low-dose anti-epileptic medication including sodium valproate and phenytoin.

    6. Vinpocetine 1 to 5 mg (the upper value seems to upset my 4 1/2 years and 17 kg son so I keep it lower).
      You mentioned Piracetam, I use Noopept and its effect is the most remarkable of all supplements I've ever tried (attention and focus) but it makes him sleep less than the usual 8 hours (witch is family disruptive but points to its antioxidant activity).
      João Santos

    7. What dose of Piracetam did you find effective?

      Vinpocetine is also seen as nootropic. The epilepsy dose is very much higher 2 to 10 mg/kg

    8. You misunderstood my words. Never tried Piracetam but its structural analog Noopept.
      I tried the 2.5 mg to 15 mg range and all values were effective (15 mg being a capsule).
      Noopept is a lot stronger than Piracetam and dosing is very different.
      João Santos

    9. Sorry João, I thought Noopept was a brand of Piracetam. Noopept seems to be the most potent nootropic drug. It is claimed that it can remove excess calcium from the brain; but this might be a misinterpretation of the actual science, which often is the case.

      " It is able to flush away excessive levels of calcium and other potentially toxic substances within the brain"

      Since excess physical calcium is a feature found in brain samples of people with ASD, this could be important, if it was actually true.

  2. My child (weighing about 60 lbs/27 kg was put on a form of valproic acid at 250 mg. His reaction to the drug was quite the opposite. It made him extremely moody and dysregulated so we stopped. We then tried bumetanide (originally 1mg in morning) and have had great success with that. We are currently dosing .75 mg in the morning, .5 mg in afternoon and .75mg in early evening and find that spreading the 2mg out has made even more improvements.

    I’m wondering if like you said in the quoted passage below he would have a positive reaction to low dose valproic acid now that we are making the GABA more inhibitory with the bumetanide. Do you know the timing of the dosage the reader is using for the valproate? I’m thinking 30mg (based on his weight) in the morning for my child but is it best to give that all at once or spread it through out the day?

    "One regular reader of this blog has pointed out that a tiny dose of Valproate, when combined with Bumetanide, appeared to have a significant and positive effect. We know that bumetanide works via NKCC1 and the GABAA receptor to make GABA more inhibitory.

    Many modes of action are proposed for Valproate, but the most mentioned one is that it increases GABA “turnover”; so it would make sense that having shifted the balance from excitatory to inhibitory, a stimulation to increase GABA signaling might be beneficial."

    As always, deep appreciation for your blog!

    1. I do not know when she gave the dose. I tried 30 mg and there was an effect quite quickly after the first dose. Giving around breakfast would seem a good time and then see if the effect lasts till bed time.

      What matters is the half life and how wide is the band of concentration where the desired low-dose effect is present.

      For Valproate half-life is 9–18 hours.

      For Clonazepam the half life is 18–50 hours.

      This means that at any time you have the effect of previous days' doses present in the body.

      Also Valproate is available as a liquid so you can accurately give a small dose.

      There is also a question in the back of my mind about irreversible positive allosteric modulation of the GABAa receptor. This is mentioned in the literature and would be a good thing if it was possible. This would occur if after giving 30 mg a day for a few months, you could stop and keep the effect. You can only find this out by trying.

      Large doses of all these benzo-type drugs looks like a very bad idea. The fact that your son had the reverse reaction perfectly fits with the original hypothesis that in some people with ASD, GABA got "stuck" in the excitatory mode. The large dose of Valproate then just amplified this excited state and made him very moody.

    2. Anonymous, I guess you ask about my son (currently 21 kg, 7 yo). I can see these social effects in him with 100 mg prolonged release formula given in the evening as well as when I give him 20 mg liquid form at bedtime and 5 mg in the morning. It seems similar to the dose range described by australian paper highlighted here in the recent post. I haven’t established the lowest effective dose for my son yet. He’s on bumetanide 2x0,5 mg.

      I didn’t see anything like that when my son was on 500 mg valproate before (with or without bumetanide). But he tolerated this drug well. He always tested low for valproate blood concentration then. It’s possible that he metabolises this drug quickly - that’s what our neuro suggested. Now he receives a kind of poly-pill as well, including some drugs that may influence valproate activity in him. So I think that the dose may need to be adjusted individually for each kid.

      I gave valproate to my son in the evening as my goal in fact was to correct his abnormal sleep EEG. I found some papers (and patents) pointing to antiepileptic use of such combinations. I haven’t checked yet my son’s EEG on these drugs.

      Larger doses of valproate may have protected my son from severe headaches associated with mast cell activation, what he is now diagnosed with as a comorbid condition. Unfortunately low doses can’t help that, he’s just had another severe flare with systemic symptoms. Peter, did you give your son liquid valproate? I am trying to find a brand which is free from additives able to induce allergic reactions, to keep my son off anything that can potentially trigger mast cell reactions.

    3. I gave Eftil liquid by Sanofi Aventis to see if there was an effect. I stopped the very low dose Clonazepam. It is very subjective, but I think the effect is not exactly the same. It is much easy to find an effective low dose with Valproate than Clonazepam.

    4. Thank you. I couldn't find Eftil online, but it seems that all liquid preparations are similar.

      I planned to compare valproate to clonazepam later on, but the headaches and mast cell symptoms got very severe in my son and we've been struggling with this for the last two weeks - too long...

      What is the difference you can see between the effects these two drugs?
      Low dose valproate apparently makes my son more affectionate and he has more pleasure in social contacts with other people. It's nice to see how this can improve relations.

      I wonder what is the role of bumetanide in all that, as the mice treated with clonazepam did not receive bumetanide nor the australian psychiatrist's adults on low dose valproate. Did he meet only patients without NKCC1 problems? I also remember that dr Chez reported best behavioral results (perhaps different than what I can see now) with higher than typical therapeutic valproate blood concentrations. This is quite complicated for me.
      Large doses of valproate and/or food allergens removal abated headaches in my son for several months and now he suffers this again. Tried oxygen, but he is not able to breath via the mask as long as CH sufferers do to stop the pain.

      You mentioned the possibility of irreversible positive allosteric modulation of the GABAa receptor, this is very interesting. Has this been shown with valproate?

    5. Have you tried Quercetin to treat the headaches? If not, it is well worth a try; its effects are evident very quickly, so you do not need a long trial, just a few days.

      I just checked and Theoharides also mentions it for headaches.

      I tried it on myself and the first dose had an effect, unlike the other things I tested. I just suggested my wife try it for her recurring winter sinus problem, it seems to have fixed that as well.

      In my son 30 mg of valproate showed an effect within an hour. It was not a hugging effect, but a hand-holding effect. So later when we went to his favourite ice cream shop, instead of running down the hill in that direction it was "daddy hold my hand" and run down the hill.

      Due to its long half life, clonazepam is harder to deal with and find the effective dose. On one occasion there was some euphoria, but the general effect was an increase in "presence", most noticeable in the afternoon 7+ hours since his bumetanide dose.

      I think most people with ASD and ADHD have the same E/I imbalance, to varying degrees. I think targeting the NKCC1 transporter and modulation of the GABAa receptor via the benzodiazepine site are just two separate ways of addressing this.

      Conventional wisdom for benzodiazeoines is:-

      ".. despite their proven clinical anxiolytic efficacy, such compounds possess a relatively narrow window between doses that produce anxiolysis and those that cause sedation"

      At these doses the effect is purely an amplification of GABAs normal inhibitory effects.

      At tiny doses, Catterall seems to be saying, that some benzodiazepines are not acting to amplify GABA's effect rather they are inverting it.

      Most likely our understanding of this area is still "emerging" and so we will not be able to figure it out 100%. In his mouse experiments Catterall also tried different benzodiazepines and the results varied. There are subunits of the GABAa receptor and different drugs have different effects on each one. It is a case where a little trial and error is needed, and since the doses are tiny this should not be problematic.

  3. Thank you for extensive answer.

    Very interesting: I can also see hand-holding effect in my son on low dose Valproate. I’ve checked our notes: it’s mentioned as remarkable behavior on several days since we started this treatment. Hugging and kissing were just more surprising to me.

    I had a bottle of Neuroprotek so I tried to use this, but my son doesn’t like its taste and I was not able to give it on regular basis, so can’t say anything conclusive. Can you share what brand of Quercetine do you use? I am not so sure about supplements quality after I was given a pack of Melatonin, which did not make me sleepy at all (fortunately other brands still do).

    1. I started with Twinlab which was in pharmacies. I also use Swansons, which you can buy online and is cheaper.

      I think some supplements are easier to produce/store correctly than others. Quercetin looks like one of those. NAC is not.

    2. Thank you, I’ll try it.

      Verapamil given just before bedtime in the late evening lets him sleep without being awaken by the pain at 3-4 AM for the last few nights. When we forgot to give the pill, this problem returned, so it’s not a coincidence.

  4. Peter - what are your thoughts on this study published in Nov '17 showing clonazepam is helpful in yet another mouse model of autism, ARID1B:
    Perhaps you have already reflected on this elsewhere?

    1. Lattegirl, as you guessed, I did cover this in another post:-

      Under-expression (Haploinsufficiency) of ARID1B in Autism and Corpus Callosum Abnormalities

      It is interesting, because it shows again that autism caused by entirely different factors (genes) can respond to the same treatment.

      It also means that people with Corpus Callosum abnormalities, that show up clearly on an MRI scan, should check for defects in the ARID1B gene and then give low dose clonazepam a try.

      I think it is very encouraging, showing light at the end of the tunnel for many people.


Post a comment