Thursday, 18 July 2019

Azosemide in Autism – ça marche aussi / it works too

Rathaus/City Hall in Hanover, Germany      
Attribution: Thomas Wolf,

The short version of this post is that the old German diuretic Azosemide delivers the same autism benefit as the popular diuretic Bumetanide, but it has a different profile of diuresis.  Azosemide may indeed be more potent at blocking NKCC1 in the brain, but this needs to be investigated/confirmed.  For some people Azosemide will be a better choice than Bumetanide.

The bulk of today’s post is really likely to be of interest only to bumetanide users and the French and German bumetanide researchers.

I did suggest recently when I published version 5 of Monty’s PolyPill, that it is getting close to the final version.  Some of the potential remaining elements have already been written about in this blog, but I have not finished evaluating them.  Azosemide falls into this category.

One theme within this blog has been to increase the “autism effect” of Bumetanide, which was the first pharmaceutical intervention going back to 2012.  I did look at modifying how the body excretes Bumetanide to increase its plasma concentration using an OAT3 inhibitor, but that is little different to just increasing the dose. There are other ways to lower chloride levels within neurons than blocking NKCC1, you can target the AE3 exchanger for example with another diuretic called Diamox, or you can just substitute bromide ions for chloride ions, using potassium bromide. Bromide is used to treat Dravet Syndrome and other hard to treat types of pediatric epilepsy.

Researchers in Germany have developed modified versions (prodrugs) of Bumetanide that better cross the blood brain barrier; one interesting example is called BUM5.  Prodrugs are out of favour because they are hard to control, meaning that they work differently in different people.

The researchers in Hanover, Germany also published data showing that an old German diuretic called Azosemide might be much more potent than bumetanide inside the brain.

This becomes even more interesting because, not-surprisingly, diuretics as drugs are produced based on their diuretic effect.  The diuresis comes from their effect on a transporter called NKCC2, but the autism effect comes from blocking the very similar transporter NKCC1 in the brain. Because Azosemide and indeed Furosemide are 40 times weaker than Bumetanide at blocking NKCC2, the pills are made as Bumetanide 1mg, but Furosemide 40mg. Azosemide is now only used in parts of Asia, where people tend to be smaller and so there are 30mg tablets (the equivalent of Bumetanide 2mg is Azosemide 60mg in smaller adults).

Then comes bio-availability, which is how much of the pill you swallow makes it into your bloodstream. Bumetanide is very well absorbed, but in the case of Azosemide it can be 20%. I was informed that you can increase this 20% by taking it with Ascorbic acid, otherwise known as vitamin C.  

In the test tube, Azosemide is 4 times more potent at blocking NKCC1 than bumetanide at the same dose.

In the test tube 60 mg of Azosemide should be very much more potent than 2mg of Bumetanide at blocking the NKCC1 transporter found in the brain.

But then we do have the blood brain barrier that seems to block 99% of bumetanide form getting through. Azosemide will also struggle to cross the blood brain barrier (BBB). The Germans think that Bumetanide is much more acidic than Azosemide and that suggests that Azosemide might be more able to cross the BBB; however the French disagree.

The conclusion of all that is to take Azosemide with orange juice.

French Researchers

You might think the French researchers at Neurochloré would have trialed Azosemide before spending millions of dollars/euros approving Bumetanide for autism.  Their patent covers all these drugs, but they would find monetizing their idea much easier with Azosemide. Bumetanide is a cheap generic drug widely available across the world. Azosemide is currently only available in some parts of Asia.

I did ask the researchers a while back if anyone had tried Azosemide for autism. The answer was no.

I think the main plan all along was to develop a more potent drug than bumetanide, without diuresis, that could be used in many neurological disorders that feature disturbed chloride levels.  The licensing of Bumetanide for autism is just an intermediate step.

There are many considerations in developing the new drug, not least what exactly is bumetanide’s mode of action. Is it the central effect of the tiny 1% that can cross the blood brain barrier? Or is it a peripheral effect?

While the German researchers think Azosemide can cross the blood brain barrier better than Bumetanide, the French do not think so.

The fact that Azosemide does have the same “autism effect” as bumetanide may help understand how it works and then this would help develop the new tailor-made drug. This is why they were interested by the news in today’s post.

I did suggest making an experiment of bumetanide and Azosemide in healthy adults to measure how much is present in spinal fluid, this is a proxy for how much is inside the brain.

In the meantime bumetanide-responders with autism have the choice of two drugs, with quite different patterns of diuresis. So for one person Bumetanide might be best, in another Azosemide and in some a combination of both drugs might be best.

Bumetanide is short-acting and causes diuresis in the first 30-90 minutes, in most people it is substantial diuresis while in some people it is minimal. Azosemide is a long-acting diuretic and the peak effect is 3 to 5 hours after taking the drug. It seems that in some people the diuretic effect is very mild and it is always delayed.
When I took Azosemide to check the effect, I did not notice any diuretic effect.  I would not have known it was a diuretic.

The higher the dose of Bumetanide/Azosemide the greater the autism benefit will be, depending on how elevated the initial chloride level was. The limiting factor is diuresis and at extreme levels ototoxicity. Very high doses of loop diuretics can damage your ears – ototoxicity.

In immature neurons you have almost exclusively NKCC1 (green above) whereas in adult neurons you have almost exclusively KCC2 (orange above), but you can be at any point in between. Also this point is not fixed in one person; external factors can shift it in either direction.

As a result the effective dose of Bumetanide/Azosemide will vary from person to person AND vary over time.

The severity of diuresis limits the dosage. This is why Azosemide clearly has a role to play at least for some people.

Here is the German paper that prompted the interest in Azosemide:-

Azosemide was the most potent NKCC1 inhibitor (IC50s 0.246 µM for hNKCC1A and 0.197 µM for NKCC1B), being about 4-times more potent than bumetanide. 

Azosemide was the most potent inhibitor of hNKCC1, inhibiting both splice variants with about the same efficacy. Azosemide lacks the carboxylic group of the 5-sulfamoylbenzoic acid derivatives (Fig. 1), demonstrating that this carboxylic group is not needed for potent inhibition of NKCC1. Clinically, Azosemide has about the same diuretic potency as furosemide, but both drugs are clearly less potent than bumetanide30, so the high potency of Azosemide to inhibit the hNKCC1 splice variants was unexpected. In contrast to the short-acting diuretic bumetanide, the long-acting Azosemide is not a carboxylic acid, so that its tissue distribution should not be restricted by a high ionization rate. However, it is highly bound to plasma proteins31, which might limit its penetration into the brain. Indeed, in a study in which the tissue distribution of Azosemide was determined 30 min following i.v. administration of 20 mg/kg in rats, brain levels were below detection limits (0.05 µg/g32).

In conclusion, the main findings of the present study on structure-activity analyses of 10 chemically diverse diuretics are that (1) none of the examined compounds were significantly more effective to inhibit NKCC1B than NKCC1A, and (2) Azosemide was more potent than any other diuretic, including bumetanide, to inhibit the two NKCC1 variants. The latter finding is particularly interesting because, in contrast to bumetanide, which is a relatively strong acid (pKa = 3.6), Azosemide is not acidic (pKa = 7.38), which should avour its tissue distribution by passive diffusion. Lipophilicity (logP) of the two drugs is in the same range (2.38 for Azosemide vs. 2.7 for bumetanide). Furthermore, Azosemide has a longer duration of action than bumetanide, which results in superior clinical efficacy26 and may be an important advantage for treatment of brain diseases with abnormal cellular chloride homeostasis.

Bumetanide in use

In 2012 I started bumetanide use at 1mg once a day and after 10 day saw a positive effect. Later I tried 0.5mg twice a day and felt the effect was much reduced.  This is not really a surprise and is highly relevant.

In the later years I increased the dose to 2mg once a day initially to combat the summertime loss of effect due to allergy (inflammation) shifting the balance of NKKC1/KCC2 further towards NKCC1.

Adding a second daily dose of 1mg produced more diuresis but no noticeable benefit. I did not try a second daily dose of 2mg because I did not want yet more diuresis.

Azosemide in use

Azosemide is a so-called long acting diuretic, whereas as Bumetanide is short acting. In practise this means there is no immediate diuresis soon after taking the drug, the diuresis comes later and can be much less. The diuretic response seems to vary widely between people.

The milder diuretic effect is attractive for the second daily dose.

After 6 years the early morning diuresis has become a normal process, but once a day is really enough. So my initial trial was Azosemide in the afternoon, while retaining bumetanide in the morning.

After a week or so there were clear signs that benefits initially enjoyed from Bumetanide have been further extended.  This is exactly as the German research suggested might occur.

After a few weeks of 2mg Bumetanide at 7am and 60mg Azosemide at 4pm I moved on to Azosemide 60mg twice a day.

Is Azosemide 60 mg more potent than Bumetanide 2mg?  It is early days, but quite possibly it is.

Bumetanide is very cheap and we have got used to the early morning diuresis, so I am less bothered with the 7am drug.

After a few years drinking a lot of water, to compensate for the diuresis of bumetanide, has become a habit. So switching from Bumetanide to Azosemide does not stop diuresis, just the urgency.

In future-users going straight to Azosemide might be a good choice.

In our case it means that a potent second daily dose is a very practical option.

Anecdotal changes include:-

Very appropriate use of bad language while driving. We live in a country with some aggressive drivers and Monty hears many people’s verbal responses to this.  Now Monty makes the comments for us.  Everyone noticed and big brother was particularly impressed.

“Car’s coming!” while extracting my car from being boxed in by three other cars in a car park, Monty noticed another car coming towards us. For the first time ever Monty has given me a loud verbal warning of danger.  He has since repeated this.  I have long wondered how a person with severe autism can ever safely drive a car, because they lack situational awareness. Many people with severe autism never learn to safely cross a road on foot.

Monty improved use of his second language. He is declining nouns and translating out loud captions and phrases he sees in cartoons.

One area I hoped would improve was at the dentist. Back in March, before the summer allergy season, we had excellent behaviour at the dentist. This gradually changed and the dentist noted this.  We are slowing repairing 2 teeth without removing the nerves and this requires visits every 7 weeks to gradually remove the decay and grow a new layer of dentine above the nerve. After Azosemide the recent anxiety disappeared and Monty’s behaviour at the dentist went back to being very cheerful and entirely cooperative.  

How to access Azosemide tablets

Thanks to our doctor reader Rene, we know that you can order Japanese drugs in specialist “international pharmacies” in Germany with a valid prescription from any European country.

So all you need is a prescription and the money.

Azosemide is available in Japan as a branded product DIART and as a cheaper generic sold as Azosemide.

The price does vary on which pharmacy you approach in Germany, one pharmacy offers these prices:-

100 Tablets   ~ 74€
           500 Tablets   ~ 286€
         1000 Tablets  ~ 524€

This is much more expensive than generic Bumetanide, but less expensive than many supplements people are buying.

If you live in North America you would have to find a different method, or take a trip to Germany.


Azosemide is still “under investigation”, but the prospects look good.

As with Bumetanide, it was approved as a drug a few decades ago and so there is a great deal of safety information. It is not an experimental drug; we are just looking at repurposing it for autism and other neurological conditions with elevated chloride.

Azosemide for autism is a good example of parent cooperation and self-help. Several parents have helped in this step forward for autism treatment.

More work has to be done to see how others respond and what the most effective dosage is.

I suspect that the optimal treatment will be twice a day and the lack of substantial diuresis in most people makes it more practical than Bumetanide twice a day.  Combining Bumetanide, a short acting diuretic, with Azosemide, a long acting diuretic, is also an option to explore.

The potential risk factors are the same as Bumetanide, disturbed electrolytes, dehydration and at very high doses ototoxicity. Ototoxicity is damage to your ear that can be caused by drugs that include diuretics at very large doses.

Azosemide would appear to have milder side effects than Bumetanide.


  1. Hi Peter, do you notice that the improvement in your son is due to azosemide rather than clemastine? I started with clemastine 1mg but don't know if the dose is too small.

    1. Valentina, I started Clemastine a long time ago. These recent improvements are just like Bumetanide ++. The alternative would be to just use more Bumetanide.

      I think most people who see an effect from Clemastine finds it takes a few weeks. I think 1mg is likely a large enough dose.

  2. Wow, this post has been cooking for a long time with many ingredients brought from readers. I'm so excited to finally read it!
    Congratulations to improved bad language and other progress! ;-)
    Peter, your meticulous work trying and documenting different dosages at different times of the day also deserves some credit, these detailed reports are invaluable for parents.


  3. Hi Peter,

    I'm so glad you're seeing continued improvement! I'll take verbal improvement with some "choice" words thrown in there any day :)

    Peter, speaking of chloride homeostasis (or disturbed homeostasis), what do you think of the following paper?

    As you know, I'm a big believer that most ASD cases are genetic (even if a genetic cause can't be identified yet, especially if its in the non-coding region) but I do believe there will be some cases that will turn out to be non-genetic. I wonder if the attached may explain at least some, and even the gender distribution, and also improvements in a large enough test population of bumetanide.

    It may be another piece of the puzzle.

    Hope all is well with you Peter, and also to friends in the community!


    1. AJ, the paper again shows the dangers caused by inflammation early in life and that females have some in-built protection.

      Inflammation in the brain is a feature of many genetic disorders. There is a cascade of effects and this means that you can potentially intervene at multiple points along that cascade.

  4. Peter,

    Have you checked out there is an app you have to download for this. Once downloaded if you type azosemide in the search engine you are able to buy direct from manufacturers lab.

    Take a look and let me know what you think I am considering purchasing not sure how much to buy for approx a 3 mth supply to try as it comes in powder.

    Hope its good product and helps.


    1. Clare, you can buy just about any chemical/drug from China, but it is best to buy the pharmaceutical product, since you then know 99% it is what it claims to be.

      It is best to find an open-minded doctor to help you. Our reader Liz in Australia recently told us that it took her 5 years to find one, but she is glad she did. It seems that some psychiatrists are particularly helpful.

      I think you need to find a doctor like Agnieszka or Rene. In the UK your NHS doctor is discouraged from off-label prescribing and all autism drugs are currently off-label.

  5. Thanks Peter,

    I am sure you are right. I am struggling to even find bumetanide now without a prescription. The first batch I brought was from mexico seller on Ebay whose page is no longer selling. The only other website I could find for bumetanide without prescription was for bumex miccil bumetanide which is the same brand I prev got off Ebay.

    Not great having to buy in such a way so yes I will look into finding an appropriate doctor.

    We are at the 20 day mark. I dont notice yet any obvious change but maybe incidents where I think short play and eye contact episode may have been caused by bumetanide. Would be great to see that improvement in eye contact sustained throughout the day. He is doing great at showing he knows what has been said to him the picking something up or putting things in the bin when told or going to the door when we talk about going out,ect. He is mimicking people but still using his peripheral vision or avoiding eye contact mostly.

    Obviously, I will give it longer but feels like I might not get a clear response I cant really imagine him changing that drastically even though I hope for him this can improve.

    Not sure what to do if he is not clearly responding? I guess I would up dose to 1.5mg x 2 daily for a while and then 2mg x 2 daily and see what happens. I know i would have to give more fluids but would i have to increase potassium and mag if i increased bumetanide??

    Thanks Clare


    1. Clare, the amount of potassium lost is a function of how much fluids are lost. The more excessive urination, the more potassium you have to add back.

      The researchers suggest trying for 3 months. I also think the changes may be subtle in young kids. If you are teaching a child addition/subtraction and suddenly they get smarter, you will notice, but in a much younger child it may be more difficult to judge.

  6. Hi Clare,

    I don't know how old is your son, but I really think it is a good idea to check blood potassium if you consider increasing the Bumetanide dose. One thing is just safety, but also some good effects of Bumetanide may be missed if a child's potassium level is not optimal. It is a very cheap test if you use private lab.

    My son has been on Bumetanide since he was 7 and indeed at the beginning the benefits were subtle so that the other Peter's blog reader saw my personal notes and concluded that Bumetanide effects are not really significant.
    Over the years all interventions targeting chloride neuronal regulation helped my son to learn many things. This would not be possible without treatment, he started with moderate ID diagnosis as well.

    I wonder if Azosemide effect on potassium loss is similar to Bumetanide.

  7. I was reading a bit about some neuroscience stuff concerning the cerebellum and Purkinje cells and came across an idea mentioned in one of the papers I was reading called "glial tonic inhibition". So I looked into this topic and found the idea very interesting and possibly another vector to look into with respect to autism. Here is one recent paper concerning the subject:


    There are many links to immune support cells in the body and the brain being compromised in critical periods in-utero for various reasons (low-grade inflammation from obesity, viral infection, maternal antibodies, etc.) and autism. Also of course excessive serotonin from the mother (usually a sign of peripheral inflammation) can change the immune profile of the baby as well.

    There has been plenty of research into brain immune cells, especially microglia, possibly having a deleterious effect on brain function by either not pruning synapses effectively (autism) or pruning too many for too long (schizophrenia), but this is the first time I have read about glial cells having a direct inhibitory effect via GABA synthesis and release directly onto GABA receptors (the paper above concerns GABA-A receptors and granule cells which are exclusive to the cerebellum and also happen to be the most abundant cell type in the brain by number).

    Peter have you come across "glial tonic inhibition" before because I have not, and if you have are you aware of any autism research being done currently that investigates "glial tonic inhibition" with respect to autism.

    1. Tyler, if you look up schizophrenia rather than autism you will find some references to glial tonic inhibition, but not much.

  8. Peter have you now dropped bumetanide completed from the poly pill and are just giving azosemide twice a day or combination of both .

    PS their is one online pharmacy that carries azosemide without prescription

    Can't guarantee this �� percent legit but this pharmacy is considered one of the more trusted online pharmacy

    1. I am still looking at what works best in our case. Both Bumetanide and Azosemide "work". The optimal therapy may vary from person to person, because the diuresis varies from person to person.

      That is the only online pharmacy selling Japanese drugs, last time I looked they did not have Azosemide. They have the 50% cheaper generic as well as the more expensive Diart version.

  9. Have you already studied about Vasopressin and regulation of testorerone precurssors?

    1. Peter has discussed both in length in previous blog posts.

      Go to Google and type this as it should help narrow your search:

      vasopressin ""

      progesterone ""

  10. A very interesting study on dopamine issues and autism came out today:

    Press Release:


    The study centered around a de novo mutation in a child with autism whereby the dopamine transporter protein (DAT) was mutated in such a way as that it both failed to do its primary job properly (recylcing dopamine by clearing it from the extracellular space and returning it to the pre-synaptic neuron), as well as aberrantly pumping intracellular dopamine from the presynaptic neuron, rather than pumping extracellular dopamine back into the presynaptic neuron. Both problems promote hyperdopaminergia at the synapse for both the pre-synaptic neuron and post-synaptic neuron, not to mention the pre-synaptic neuron runs out of dopamine for phasic release. On top of that, the increased levels of extracellular dopamine desensitize D2R receptors which leads to decreased dopamine synthesis in the presynaptic neuron, thereby decreasing proper phasic dopamine signaling and instead a situation where you have a noisy dopamine signal from the excess dopamine in the extracellular space with no phasic dopamine bursts coming from intracellular pre-synaptic sources, thereby leading to big deficits in reward learning.

    Now this is just one de novo mutation, but the really important thing here is the researchers found a gene in a human that caused autism and then put it into a mouse and produced autism symptoms, rather than the typical situation where researchers toy around with a bunch of genes in mice and find one that causes "autism symptoms", but then often does not do anything substantial in humans.

    Last but not least, the researchers improved symptoms in the mutant mice with a DAT blocker (not to be confused with a DAT Reuptake Inhibitor which in effect does the opposite of what would be intended with a DAT blocker).

    Therapeutically, reducing excessive extracellular dopamine levels via BCAA's or other therapies to reduce dopamine synthesis in the brain may solve half of the problem with this specific de novo mutation, but it likely won't restore phasic dopamine release because the mutant DAT protein will still be pulling dopamine out of the pre-synaptic neuron, rather than pushing dopamine back in as it is supposed to function. In that case I suppose you would need to increase dopamine levels somehow in the pre-synaptic neuron, while decreasing extracellular dopamine which makes BCAA therapy have the problem of both reducing extracellular as well as intracellular presynaptic neuron dopamine (which predominantly comes from the substantia nigra).

    1. Hi Tyler, hope your sons are doing well, thanks for this study about dopamine. As DAT blockers are strong prescription meds, could the problem be solved by taking BCAAs with something that increases intracelular presynaptic dopamine? And what could be?

    2. Tyler, your excellent comment and this new study has mixed my day(s) completely...

      Just one little correction - with specific DAT blocker / ACT-01 / they were able to improve only one symptom in mutated mice / not symptoms / and it was hyperlocomotion. Other social and repetitive behaviors were not improved. And they conclude that it was done only by preventing the DA efflux through the mutated transporter into the synapse.

      This study is very important for me and my son, because we have done 3 years ago genetic testing through 23andme and he has a mutation in the SLC6A3 gene / rs28363168 /, which codes the dopamine transporter. I just was not able to find if his mutation is one the known mutations / besides the A559V / rs28364997 / which is linked to ADHD - so this one he does not have for sure /. There are no other rs numbers to least I was not able to find any.

      After we found out his SLC6A3 mutation we have tried RITALIN / methylphenidate, Dopamine reuptake inhibitor / and had to stop it after 6 days. It caused worsening in almost every autistic symptom. As you wrote Tyler - at that time we probably did the opposite...

      We tried for several months also the ASD popular ketogenic diet and it ended in our case as disaster too...

      Although this failed interventions were very hard to digest for us, they gave us now extremely valuable informations.

      The same group of researchers did 2015 also this study :

      and they found out in vitro that the functions of mutated DAT transporter - in the case of Y355A mutation DA uptake and in the case of T356M mutation both DA forward and reverse transport, were partially restored with Zinc.

      In my eyes the best cast scenario would be if his mutation would cause just a defect of Dopamine Re-Uptake from the synapse. In my current understanding the defect in Dopamine efflux into the synapse is more associated with ADHD symptoms and these he definitely does not have.

      The too high brain dopamine levels could be than targeted by BCAAs therapy, zinc supplementation and by partially restricted dietary intake of high proteine foods. Thats my theory right now.

      Thanks for every opinion !!!

    3. Was there any particular type of zinc that improved symptoms, or was any form of easily digestible zinc recommended?

    4. No, there was not specified type of zinc supplement.

      Another excellent paper about DAT from danish group, which is collaborating with US researchers from the latest DAT paper :

      There is also a lot about Zinc effects on DAT. Interestingly, in WT-DAT Zinc acts in a similar way like Ritalin, as reuptake inhibitor. But in some mutated DAT effect change from inhibitory to activatory, and therefore is Zinc partially restoring defect functions of mutated DAT. Problema is that it does not happen in every mutation, only in some of them - so you have to know exactly which one you have and also if this mutation was previously tested on Zinc effect.

      I am in contact with both reseacher groups right now / US and Denmark /, lets see what I can find out.

  11. Are there any natural diuretic supplements that have similar effects in balancing NKKC1/KCC2 towards NKCC1 for those who have difficulty obtaining a prescription/ordering online?

    1. Even the prescription drugs have only a modest impact on NKCC1/KCC2. In the lab some natural substances do have some effect, but in reality I think they will have a trivial impact. One is Ginseng.

      In some people with severe autism lowering chloride levels has a profound impact shifting them away from profound MR/ID. So I think it is well worth the struggle to obtain bumetanide or azosemide, just in case the person is a responder.

  12. Hi Peter
    I am 3 weeks into trial on my 4 yr old son and 30mg Azosemide looks good so far and no duresis at all.
    Can see some subtle changes but will trial for 3 months and then see if it's working and if there are any significant gains..
    Is it a good idea to give one 30 mg dose or split in twice a day ?


    1. Vicky, the experiences I have had to date suggest it is better to have a larger single daily dose than to have 2 small doses.

      For most people the only side effects of azosemide/bumetanide are ones causes by the diuresis. So it is the diuresis that limits the dosage you can give.

      In my son the more bumetanide/azosemide the better/bigger the effect is.


Post a comment