Thinking Person's Guide to Autism

I wrote this list because most people visit this blog and do not go any further into the details, either because they think this blog is nonsense, or it is just too complicated.

It certainly does sometimes get complicated, but complex subjects require detailed attention. There is little point in looking superficially.  Understanding autism science is not beyond most people’s ability, but it does take time. It will not cost you a penny. Then you can make up your own mind as what is a novel therapy and what is likely a quack therapy.

Some doctors do know best.  While your family doctor no doubt was taught autism is untreatable, some doctors who have a child with autism think differently and I do not mean MAPS/DAN doctors. Some mainstream doctors read the same research as me and successfully treat their child – an example of personalized medicine. In the ideal world they would all publish their experiences, but most keep it private.

Autism is not a biological medical diagnosis, it is just an observational diagnosis based on guidelines published in a book called the Diagnostic and Statistical Manual of Mental Disorders (DSM). These guidelines have changed substantially over time. We are currently on DSM 5; in 1987 it was DSM3 Revised. Most people diagnosed today with autism would not have been given that diagnosis under DSM 3. This is one key reason for the autism diagnosis “epidemic”, but there are other reasons.

A vast amount of scientific research already exists about autism. More is published every day and very helpful lay summaries are available for free on the internet on medical news websites and blogs like Paul Whiteley's Questioning Answers. There is also the Simon’s Foundation’s excellent website for non-scientists at

There are hundreds of different known biological causes of autism.  Scientists have already created 267 different genetic mouse models of autism and 78 types of induced autism. 192 rescue models exist, where the scientist could reverse autism in the mouse.  If you can reverse it in a mouse, perhaps you can treat it in a child? Not crazy to at least try.

Some autism/MR/ID caused by rare errors of the metabolism is treatable today, by mainstream medicine. Clinicians in Vancouver maintain a database; they focus on Mental Retardation (MR) / Intellectual Disability (ID), but many of these conditions can also be diagnosed as autism.

Science and medicine are not the same. Medicine applies drugs that are shown effective in large clinical trials on people with the same biological disorder.  This is evidence-based medicine.  Where a disorder exists with multiple causes, medicine struggles to cope. Many illnesses have multiple causes, or unknown causes, for example dementia, depression, epilepsy, multiple sclerosis etc. and even conditions outside the brain like chronic prostatitis.

Medicine has many poorly effective drugs for neurological conditions. There are drugs for dementia, ALS (motor neuron disease), Parkinson’s, Huntington’s, Schizophrenia, Bipolar etc., but they are not curative - they are better than nothing, but sometimes not by much. Drugs for autism should be seen in this perspective; generally they will be partially effective and only in specific people with the same particular biological dysfunction. Drugs for ALS, Alzheimer’s etc. currently just treat some features of the disease, they do not address the cause; some of these features are shared by others brain disorders. So using an ALS drug to treat autism or schizophrenia is not a crazy idea, if the same biological features are present.

There are numerous well-researched biological features present in some autism. Some of these same features are present in unrelated, better-researched, medical conditions where often they are treated. For example:- oxidative stress, elevated pro-inflammatory cytokines and reduced anti-inflammatory cytokines, activated microglia, disturbed growth factors (BDNF, NGF, IGF-1, VEGF etc.), numerous ion channel and transporter dysfunctions (NKCC1, KCC2, Cav1.2 etc.), NMDA hypofunction or hyperfunction, reduced estrogen and ERβ, excitatory-inhibitory imbalance, glutamate excitotoxicity, impaired autophagy, unusual myelination etc. 

Autistic brains are not just “wired up differently”. Brains are not computers and while some things are fixed as in a computer, much inside your brain is changing all time and so can potentially be modified.  You can induce even “autism” in a perfectly normal brain (with propionic acid) and then reverse it.  A slightly better analogy might be an old out-of-tune piano; tuning the strings (wires) is a complex process, but results in much improved function. Autistic brains can be tuned to improve their function and the person’s wellbeing.

Vaccines can trigger mitochondrial disease in children, which will appear as autism. This was proved by Dr Jon Poling a US neurologist trained at Johns Hopkins, who won $1.5 million in compensation for his daughter. Here interviewed on CNN:-

Vaccines caused one little girl's autism, but that does not mean they caused it in your child. How widespread this problem might be is something Dr Poling does comment on, but it is not something anybody would research.  Johns Hopkins is home to some of the best autism and mitochondrial disease researchers and clinicians. They even developed a cheap therapy to minimize the risk of possible damage from vaccinations to their young patients who already have mitochondrial disease. 

Autism does not kill you?
Vaccines save millions of lives and even Dr Poling does not suggest avoiding them. People with severe autism do have a life expectancy of less than 40 years mainly due to death from seizures, drowning and other accidents. People with the Asperger's type of milder autism have a nine times elevated risk of suicide and a disturbingly a high proportion of the tiny number of mass killers have an Asperger's/autism diagnosis (the other large group have suffered a previous head injury). So autism clearly can have troubling consequences.  Teaching anyone to shoot a gun who has a neurological disorder (autism, bipolar, schizophrenia, even ADHD, Intermittent Explosive Disorder etc), severe enough to get a medical diagnosis, is asking for trouble.

Does Autism need to be treated?
Some of the people with the newly included mild autism protest about the idea of treating autism. A significant minority of deaf people think their children born without hearing should not receive cochlear implants and so remain with them in silence, but most hearing people would likely want their child disabled by hearing loss to be treated. Medically diagnosed autism is currently based on guidelines from a book of mental disorders (DSM5). Most people probably think disorders in a psychiatrist’s manual should be treated, even if the very people with those disorders think they are just fine - they do after all have a disorder potentially clouding their judgement. If their autism is so mild and not troubling at all, then it does not warrant a medical diagnosis. You can be different without needing a medical diagnosis to support it.

Many observational diagnoses like Autism, ADHD, Bipolar and Schizophrenia are overlapping biologically
. Genetic studies have shown people can have biological elements of more than one diagnosis. So you can have autism with a little Bipolar and a touch of ADHD.  This means that some therapies trialed in Schizophrenia may show a positive effect in some autism, for example.

Clinical trials in Autism have all ultimately failed. All of the many clinical trials in autism to date have been viewed as failures and part of the reason is that there is more than one autism. There is no easy way to tell which person has which type, so trials include many people who are bound not to be responders.

People treating autism are all quacks? There undoubtedly are many people making money out of the families affected by autism and where money is involved, you will always find quack therapies of one kind or another.

Personalized medicine is one way forward. In cancer care personalized medicine is already being used, where drugs are tailored to exactly what is wrong in a specific person.  With hundreds of different variants of autism personalized medicine would be the ideal solution, but it is many decades away for most people.

Autism clusters and nexus where multiple dysfunctions converge.  The realistic way to treat autism in the 2020s will be to group people into similar clusters, not perfect matches, and treat by cluster.  This is possible because it seems that multiple different dysfunctions converge at a manageable number of so-called nexus. This means that rather than several hundred unique therapies you have just a few tens of therapies. Each person is then matched with their Polypill of a handful of those few tens of therapies.

Autism is a condition of opposites. In autism very often both extremes of the same condition exist, making the average meaningless. On a simple level the research finds many big heads, but also a fair number of tiny heads. There is very high cholesterol and very low cholesterol. There are very high levels of a particular growth factor and then some with virtually none.   So do not expect what works for someone else’s autism, to necessarily work for your child’s autism.

Many drugs have multiple modes of action. Lay people assume that drugs do just one thing, be it lower blood pressure, lower cholesterol or act as a diuretic.  This is not the case; most common drugs have multiple effects. For example many antibiotics have completely unrelated anti-inflammatory effects. A diuretic that affects one ion channel in your kidneys can also affect a related one in your brain, so yes a diuretic can treat some people’s autism, it is not crazy.

Many existing drugs can be repurposed. Many drugs already exist that can potentially be repurposed to treat neurological conditions. Big drug firms like to develop new drugs because their patents allow them to charge very high prices. Existing drugs are often now cheap generics, with well-known safety profiles, that can potentially be used off-label in personalized medicine, based on intelligent use of the vast wealth of autism research.

Off-label prescribing used to be widespread; it is when a drug is used for an application other than that for which it has been officially approved. For example, the diuretic Spironolactone is widely used off-label to treat acne in females.  Modern doctors are less willing to prescribe off-label either because they insist on strictly following the rules, or do not want to take a risk of being blamed for an adverse reaction.  Treating autism will inevitably require off-label prescribing.

Personalized therapy for the few? Hopefully the mainstream doctors who currently successfully treat their own children will persuade more of their peers to try and treat others. Until then, or until a drug like Bumetanide finally passes its stage 3 clinical trials for autism, I do not think much will change.  A small number of parents do read the research and persuade their physician to help them and others evidently self-treat. For those who are truly motivated there are options, which is what really matters.


  1. Good afternoon Peter and readers!
    I don't have basic scientific knowledge and need so much guidance to avoid mistakes I've already made.
    My son diagnosed with ASD at 2y8m, slightly developmentally delayed, mainly speech+behavioral anomalies. Since then for a year in therapy with two US (MAPS) docs with min improvement, but remarkable improvement with ABA. Persistent gut dysbiosis, allergies, history of 5 in a row HSV in the mouth (age 2). Extensively investigated for inborn errors of metabolism by university lab here in South Africa, no conclusive diagnosis, scheduled for International metabolic forum. Mild branched chain amino acid abnormality, dysregulated immune system, suspected mito issue (raised urine tyglyglycine and 3-methylglutarate), acylcarnitine profile essentially normal, hyperopia, aniosometropia, horrible toxic profile. Brain MRI normal, but small CSF-density in splenium of corpus callosum. 24hr EEG normal. Confirmed Fragile X a few days ago. Array-CGH normal. Despite all of these a bright handsome boy with lots of energy and very-very capable.
    Peter, Roger, where do I start here?
    - Will try NAC and broccoli first thing
    - Ordered generic Leucovorin
    - went through listed here trialed drugs for FXS
    Which FXS parents forum uses science?
    And please if someone can help me find the knowledgeable doc!!! Doesn't matter where he is.
    Peter, I am in South Africa, I am fluent in Russian and Ukrainian if I can be of any help for this blog.

    With so much gratitude for all you've posted here,

    1. Mira, now you have a biological diagnosis it is much easier.

      At least some people (I would guess most) with Fragile X have the same problem with high chloride as many with classic autism.

      So bumetanide should be trialled ASAP. I think this will help your son a lot and raise his IQ.

      Treating Fragile X syndrome with the diuretic bumetanide: a case report.

    2. Mira, I would also try a statin, like I use. This is because Fragile X affects RAS and statins modify RAS signalling.

      Cholesterol drug may treat fragile X syndrome

      This is another paper about Fragile X and bumetanide

      The Developmental Switch in GABA Polarity Is Delayed in Fragile X Mice

      Also, Baclofen or the experimental drug Arbaclofen, helps some with Fragile X and there is a very similar Russian drug called Pantogam.

  2. So so appreciated! Contacted Dr. Mike Tranfaglia also

    1. The FRAX research foundation have some good ideas, but they grossly underestimate the cost of repurposing an existing drug and how long it takes (Bumetanide, Suramin etc).

      The only shortcut is experimental use, which is frowned upon.

  3. Dear Peter,
    I started bumetanide 6 days ago, at 0.5mg 2t/day now, saw a few absolutely fantastic things happening already with my FXS and ASD 3.5 yo 17 kg son. Hope it will hold! But stereotypy (verbal and visual stimming), that was always very-very little, definitely increased.
    He gets now Mg, KCl 200mg (is chloride OK? getting tablet form 99 mg KBr in a few months from Ukraine, but with Adonis Vernalis extract in it($3 for 20 tabs!), Australian broccoli sprout powder (was immediate responder to this with fluent speech increase, thank you!!!) and NAC (not sure of this one yet).
    What do you think I can do about stimming? Looks like it came with bumetanide, but sensory gate is much better with potassium now (my son refused to sit inside the trolley in the mall today! Never happened before).
    Read a few people on the blog had a similar problem.

    And thank you so much again,

    1. Mira, I think in some people who are big responders to bumetanide, where it is like the fog has finally cleared, an increase in stimming is a natural reaction. In my son NAC helps greatly with stimming, but it depends why he is stimming. I think oxidative stress causes anxiety and stimming. But other things also cause anxiety and stimming.

      NAC has a short half life and so you need to give it frequently. So you could try more NAC for a few days and see if this reduces stimming.

      The broccoli makes some people a bit hyper, which also might increase stimming.

      Since you have added several therapies, I would wait a few weeks for things to stabilize. Maybe stimming will reduce.

      You can also use behavioural strategies to redirect the stimming. You could try soft tactile toys, there are silicone stretchy ones and foam "squishies".

      KCl must be OK since you are getting a benefit from bumetanide, people also ask about eating table salt (NaCl).

  4. Dear Peter,

    3 weeks since Bumetanide+KMg start (gains in communication, socializing, bonding,awareness, sensory issues, etc.)
    With those gains extreme anxiety, hyperactivity and visual stimming (spinning, lights, water, etc.) kept growing every day.

    After 3 weeks I decided to stop Bumetanide for a week to see if anxiety and stimming will reduce. 5 days later don't see any reduction, but it definitely increased with Bumetanide+KMg.

    My son is taking Broccoli Sprouts, ALA (no effect so far), trialled NAC 4t/day - no effect also (used Solgar brand 600mg, opened each cap, tastes sour, but no smell). The rest - Quercetin, Olive leaf extract, B12 inj and probiotic - we are using for a long time.

    Had 2 consults:
    - top MAPS doc said to give 200.000 VitA for 3 days to see if visual stims decrease otherwise it might be viral issue (Rubella?). Strange

    - famous US doc and FXS researcher gave me methylphenidate script for ADHD and advised to trial Acamprosate.

    If you get a minute, please give me your thoughts on this. ABA therapists complain, can't make him concentrate on anything any more, he is running and jumping like mad, but nobody denies the gains in communication, socialization, etc.

    Toys to redirect visual stimming like kaleidoscope, glitter bottles, etc don't work at all. Give him a lot of movement as reward

    And thank you so-so much again!

    Kind regards,

    1. Mira, Acamprosate is interesting; it is increasing the effect of GABAa receptors and reducing the effect of NMDA receptors. In typical people (recovering alcoholics normally in this case) this will increase inhibition to neurons firing and reduce excitation, so generally calm things down.

      The drug is approved in the US to treat Fragile X.

      Using Bumetanide you are correcting what GABAa receptors do, rather than changing the magnitude of their effect. So I think Bumetanide should be a better therapy for Fragile X.

      I am not so sure what would happen if you use Acamprostate and Bumetanide; it might be great or might not be, so you should be cautious if you use both.

      Ritalin/methylphenidate often causes lots of problems, so I would avoid it and all other psychiatric-type drugs.

      Did you check his electrolytes levels after starting bumetanide? Perhaps these were depressed and triggered the problem?

      If it was me, I would stop every drug/supplement for two weeks, including the ones you have used for ages. Then I would reintroduce bumetanide and monitor potassium and other electrolytes. Adjust your K+ supplement to maintain it around the higher end of the reference range.

      If he again becomes hyperactive, I would continue with bumetanide to see if he adjusts over time and the hyperactivity fades.

      I would use as few drugs/supplements as possible because then you minimize the chance of unwanted interactions. Broccoli powder makes some people hyperactive. Perhaps when taking bumetanide your son becomes one of those people. B12 also causes hyperactivity in some people.

  5. Checked electrolytes before and 2 weeks after Bumetanide.
    Maintaining high normal K. Epiphanyasd is my manual.

    Exactly my thoughts about stopping everything and reintroducing Bumetanide alone. Thank you SO MUCH for going into such detail for me!

    For your ref: 3rd opinion (FRAXA research doc) says he might react behaviorally to electrolyte disturbances and urinary frequency and this is certainly not a good response as we expect behavior would improve along with cognition. He admits though he doesn't have clinical experience with Bumetanide as it seems not ready for routine use based on evidence to date.

    Thank you SO MUCH for going into such detail for me!

  6. Good afternoon Peter!
    Using your vast knowledge and research, do you think there's ever a slightest chance to prevent cognitive decline in FXS if starting to treat early? Or FMRP absence is a sentence?

    Appreciating your every word,

    1. Mira, lack of FMRP is going to have numerous effects, but some of these effects appear treatable. The sooner you start to treat the consequences of the lack of FMRP, the better the result should be. All improvement is relative to what would happen with no intervention.

      This applies to all such genetic conditions, so success in treating someone with Down syndrome might be raising IQ from 60 to 80, and one parent of a child with DS left a comment that such a change was all he hoped for. Parents have to be realistic.

  7. Dear Peter,
    We sadly dropped Bumetanide for my 4yo with FXS as we didn't see any further gains, urination went up to literally doing it every 5 min in 40-50 min after taking the drug, accidents. K and electrolytes normal.

    Want to trial Metformin and low dose Sertraline 2.5-5.0 mg (per R. Hagerman). What will be your thoughts? Can I substitute SSRI with sth else? Is there anything that can serve as a marker that my son might be a responder to Metformin or Minocycline? What would you trial first for FX? (Statin and Pantogam aktiv are on the way also).

    His blood sugar is normally low (effect of mild ketogenic diet), normal weight and homocysteine, often bloated, mito insufficiency around keto and glutaric acids.

    Another question: He has 9p24.3 copy gain with hyper-IgE recurrent infection syndrome in that region. Only dust mite allergy medically confirmed, rest negative.
    MAPS doc suggested Xolair or even IVIG. Do I have to go that far with his allergies?

    Highly appreciated


    1. Mira, if you try a statin, use Lovastatin. In the FXS mouse model Simvastatin gave no benefit, but Lovastatin corrected features and biomarkers. Lovastatin is less widely available.

      Low dose Sertraline or Fluoxetine is interesting and I have an unpublished post on the subject. At low doses, these drugs are not effective SSRIs, but they do increase the amount of the steroid hormone allopregnanolone. This in turn acts as positive allosteric modulator (PAM) of GABAa receptors.

      There may also be an effect of the expression of KCC2 receptors, which then changes chloride levels inside neurons (which bumetanide also does).

      So I think following Dr Hagerman is a great idea, particularly for FXS.

      There is solid evidence to support a trial of metformin for FXS.

      The marker that your son might respond to Metformin, Sertraline and Minocycline is his FXS diagnosis.

      I would wait and see if his behaviour changes at time of allergy. If behaviour changes then it is a good idea to consider possible therapies. If an exposure to his allergen does not impact his behaviour then I would not rush into expensive therapies (IVIG is expensive and you have to keep repeating it).

      I think you have a good chance of success with all the above. You need to try them one by one, to be sure of what is giving what effect.

    2. Hi Mira, I know this is a bit of an old thread, but I'm curious how you found a MAPS doctor who was well-versed in allergy/immunology topics. I'm trying to find someone in that category but have not had much luck with internet searches. Thanks for any suggestions.

  8. Hi Peter,

    I completely agree with your post and just got into a London accelerator based on the idea to personalise Autism interventions and treatments using big data. I'm on this mission because I have helped raise my nephew who was diagnosed with ASD and have taught many children who face the same difficulties of ASD. My family has been compiling our own trial and error log of what works with my nephew and we've used the epiphany blog for research. We really appreciate your research as it has helped us, although we still haven't figured out the right solution just yet.

    My background is in ABA therapy (9 years) -having taught children on the ASD spectrum with (5 years) of experience in social impact startups, and my cofounder, Jurek, holds a PhD in computational biology and 11 years of AI expertise. Together we're exploring solutions for personalised Autism care. We would really appreciate getting in touch with you directly to see if we could chat as you're one of the best practical researchers we've seen. I've tried getting in touch a few times and am not sure what the best method is.

    If you or anyone is interested in the work I'm conducting, please let us know. We are scouring Autism research resources to create a tool that can help parents for their children. We will be creating a free app as I know how costly Autism can be.

    My email is zeena (at)

  9. Dear Peter!

    Will you be of the opinion that LOW pH (6.9) in blood gas profile is a reason to consider HBOT in FXS? Or is it worth considering only for high pH?

    Do you think Lipostabil inj might be a good therapy for FXS?

    What will be your thoughts on why Duke's stem cell trial excludes FXS?

    Very thankful,

    1. Mira, the blood gas test also measures bicarbonate, oxygen saturation and partial pressure of O2 and CO2. Your doctor needs to look at all the results to figure out why pH is low. Then you can see how to best raise it to the normal level (7.4).

      Specific types of metabolic acidosis each have their own treatments. I would run the test again to confirm the result and go and see a good doctor. I doubt HBOT is the answer, but raising pH should be a priority.

      Lipostabil injections might help some very rare metabolic disorder, but it is not something I know anything about.
      I would focus on the existing FXS treatments in the published research.

      Clinical trials often have strange exclusions. FXS is a precise diagnosis and so if a therapy works for one person with FXS there is a good chance it will work in another person with FXS; you cannot generalize in this way for what works in someone with an autism diagnosis.

    2. Mira, this study shows how research is excluding more severe types of autism.

      Are Children Severely Affected by Autism Spectrum Disorder Underrepresented in Treatment Studies? An Analysis of the Literature

      Despite significant advances in autism research, experts have noted that children severely affected by autism spectrum disorder (ASD) appear to have been understudied. Rigorous analysis of this observation has been limited, and the representation of severity has not been well-described. We assessed three domains of severity (communication ability, cognitive functioning, and adaptive functioning) in 367 treatment studies of children with ASD published 1991–2013. We found that the proportion of studies that included the severely affected population decreased significantly over time, as well as wide variability in measurement and reporting. Inadequate representation of the full autism spectrum in the literature could lead to an unbalanced picture of ASD and leave behind those with arguably the greatest need.

  10. Hi Peter. Hopped over to your blog via the Longecity forum. Lots to explore here! Thank you for keeping up to date with emerging news.

    I'm a filmmaker with dyslexia, in the queue to been seen for suspected asd/adhd.
    I'm going to test some of the supplements listed for effectiveness here in meantime. Just wanted to say introduce myself and show my appreciation!

  11. Hi Peter, any thoughts on Piracetam for a 5 yr old girl with autism. We need help with language.

    1. guy, i used piracetam for my kid 3,5y. combine with citicolin and galantamine. no effect,but safe. maybe my kid have enough acetylcholine, maybe autism kid have problem with gaba, not cholin like alzheimer

    2. Piracetam is worth a try. It is cheap and safe.

      For language specifically, it looks like your best chance is calcium folinate. The research tells us that about 70% of people tested have antibodies that reduce the transport of folate to the brain. Such people have normal levels in their blood, but very reduced levels in the brain as measured in spinal fluid.

      Girls with severe autism are quite likely to have an identifiable genetic cause of their autism. Whole exome sequencing is a good idea particularly in girls or for anyone with dysmorphic features.


Post a comment