A good idea gone horribly awry

In most scientific disciplines, medicine included, hypotheses start with observations.  When John Snow became curious about a London cholera outbreak, he looked for patterns and found that many cases centered around a local well.  While he didn’t have a clear idea what caused cholera, he suspected the water as a source, and tested his hypothesis by having the pump handle removed, which appeared to help stem the epidemic.  Eventually it was discovered that that a bacterium was the cause of cholera and that is was transmitted in drinking water contaminated by feces.

More modest hypotheses are the grist of the science mill.   Researchers may observe a new trait in a fruit fly and hypothesize that it is due to a novel genetic change.  They then do the hard work of looking for correlations and associations, and the bench work involved in pursuing an answer.  In medicine, testing a hypothesis at the bench is followed by testing on real people, a significant scientific and ethical shift.

There are a myriad ways that medical science can go awry, and as in much of human thinking the root of many of these mistakes is our own fallible brains.  We see patterns, and when we see patters, we often assume that post hoc, ergo propter hoc, that is, if one event precedes another, it is the cause.  When you take the tinder of this human tendency, and ignite it with human emotion, we can end up consumed by medical catastrophes.

Fellow Scientopia blogger Female Computer Scientist tipped me off to a piece in The Scientist, one that demonstrates how a father’s desperation and a doctor’s desire to please led to a terrible medical decision.  The article tells the story of the Johnson’s, whose autistic son developed intractable behavior problems.  The family pursued all of the usual conventional medical approaches, purposely avoiding much of the autism quackery that so many others fall prey to.   As their son worsened, Mr. Johnson started doing his own research—and stumbled onto something interesting:

He discovered the work of a trio of physician/researchers at the University of Iowa who had successfully treated patients with Crohn’s disease and ulcerative colitis using a nematode parasite found in the intestines of pigs—Trichuris suis, the pig whipworm. Both are autoimmune disorders in which the immune system essentially attacks the intestinal walls. Stewart also found data that pointed to a link between some autism symptoms and inflated levels of proinflamma​tory cytokines, an apparent result of the immune system attacking glial cells in patients’ brains.1 Putting these bits of information together, Stewart wrote a short review paper and presented it to Hollander. His central hypothesis was that parasitic worm infection would modulate Lawrence’s immune system and calm inflammation that was causing his disruptive behaviors.

This hypothesis—that autistic behaviors can be modulated by parasitic infection—is not as insane as it sounds.  Rabies increases aggressive behavior in animals, causing them to bite and spread the virus.  The parasite Toxoplasma has been hypothesized to change human behavior. But in the Johnsons’ case, an unbroken chain of faulty assumptions led them to try something foolish.

The basic syllogism is this:

  1. Inflammatory diseases can be treated by a parasite.
  2. Autism is an inflammatory disease.
  3. Autism can be treated by a parasite.

First, can “inflammatory diseases” really be treated with pork whipworms?  The studies are unimpressive.  The few studies done found some minor statistically significant improvement when whipworm infestation was tested against placebo.   The inflammatory response generated by parasitic infection is a bit of a blunderbuss.  Compare it to something like infliximab, a monoclonal antibody very successfully used in Crohn’s disease that targets a specific immune molecule and doesn’t involve ingesting worm larvae.

Many diseases fall under the category of “inflammatory”, and inflammation is a complicated process, one that can’t be nailed down as something to be turned “on” or “off”.  This makes analogies between one inflammatory disease and another problematic.

Is autism an “inflammatory disease”?   One study cited in The Scientist looked at a very small sample of autistic people, without a control group, and found evidence of inflammation in their neural tissue.   There  is no clinical evidence that autism has an “inflammatory” cause that can be affected by immunomodulation with, say, corticosteroids which are a non-specific immunologic treatment effective in many inflammatory diseases.

Given that statements one and two are false, number three is ridiculous.  But the Johnsons were understandably desperate and tried worm therapy.  At low levels of infestation, they saw little improvement, but at high infestation levels, there was a change:

The Johnson family anxiously awaited the effects of the full dose of TSO on Lawrence’s violent behavior. Within 10 weeks of the higher-dose treatment, the autistic boy stopped smashing his head against walls. He stopped gouging at his eyes. The paralysis and frustration that held him and his family prisoners in their own home lifted. The freak outs ceased. “It wasn’t gradations,” remembers Stewart, who had always kept meticulous notes on Lawrence’s disorder and the interventions they had attempted. “It just went away. All these behaviors just disappeared.” Elated, Stewart called Lawrence’s doctor, Eric Hollander. “He was stunned, because all of that behavior set was gone,” Stewart says. “He was speechless, as I was.”

As Lawrence’s previous behaviors demonstrated, there was a certain unpredictability to his disease.  He sometimes worsened, sometimes improved, and in cases like this, we are even more easily fooled by our post hoc ergo propter hoc thinking.   There is no way of knowing, in this single case, whether or not worm infestation had a clinical effect on autism.

I suspect this story is going to continue on two separate paths. Quacks are going to take it and run, marketing expensive whipworm therapy to desperate parents.  Real scientists and clinicians are going to take a step back, and examine the possible role of inflammation in autism on the one had, the the possible role of parasites in various inflammatory diseases on the other.  The work will be hard, slow, and lead to many dead ends, but eventually we will learn important scientific facts.

The slow pace of science can be torture for those who are suffering, but turning individuals into uncontrolled experiments helps neither the patient nor science as a whole.  We treat lab rats better than that.


SUMMERS, R., ELLIOTT, D., URBANJR, J., THOMPSON, R., & WEINSTOCK, J. (2005). therapy for active ulcerative colitis: A randomized controlled trial Gastroenterology, 128 (4), 825-832 DOI: 10.1053/j.gastro.2005.01.005

MAYER, L. (2005). A novel approach to the treatment of ulcerative colitis: Is it kosher? Gastroenterology, 128 (4), 1117-1119 DOI: 10.1053/j.gastro.2005.02.038

Vargas, D., Nascimbene, C., Krishnan, C., Zimmerman, A., & Pardo, C. (2005). Neuroglial activation and neuroinflammation in the brain of patients with autism Annals of Neurology, 57 (1), 67-81 DOI: 10.1002/ana.20315


  1. I’m not a clinical scientist, but I thought the Crohn’s disease studies were pretty impressive. And there’s plenty of evidence supporting a link between worm infections and decreased autoimmunity. Though the jury is still out, I actually think your point 1 is at least plausible.

    That said, the logical leaps here were pretty surprising, and I agree completely that we shouldn’t generalize this one patient. But reading the article, I was pretty happy with the amount of hedging, and admission that far more work needs to be done. Also, I’m pretty encouraged that this parent rejected pseudoscience, generated a hypothesis and went through a licensed doctor with the approval of the FDA. It might not have been the best way to do things, but at least he didn’t go to Africa to infect his son like this guy.

  2. Jess

     /  February 10, 2011

    I second Kevin – I think worm therapy holds quite a bit of promise. And at the very least, as you suggest, controlled clinical trials are small and young, so suggesting that point 1 is “false” is a big over-reach.
    I should mention my biases: I am currently working in a neuroimmunology lab which is collaborating on a clinical trial of TSO in MS patients.
    There is a lot of basic science and epidemiology data that informs the hypothesis that worms may help people with certain pathological immune environments. There are studies of uncontrolled, naturally acquired worm infections that demonstrate a correlation between infection and milder MS disease. And there are a number of clinical studies occurring at this moment to determine the value of worm treatment in various conditions. Is it too soon to start passing out TSO like candy? Absolutely. But it is also quite premature to write it off at this point.
    Another motivation to pursue studies in worm therapy is the fact that other treatments for various immune pathologies can be really un-fun for patients and frankly may not change the disease course that much. Blanket immunosuppression, or even more specific suppression targeting certain cell types, can have serious side effects which make these treatments unacceptable to many patients. Worms at least are pretty safe – about a third of people in the world are intimate with human-specific worms, and when you’re well-fed that’s not necessarily a huge problem. TSO takes that a step further by using a worm that doesn’t spread amongst people, doesn’t reproduce in the host, and never lived inside another human.

  3. This happens to relate to my research on autism and on the hygiene hypothesis. This result is fully compatible with my conceptualization of the low NO causation of autism. The use of worms like this was discussed in the same book on the hygiene hypothesis that has the chapter on my bacteria. They all work by the same mechanism.


    My bacteria would do this too. The improvements to my Asperger’s took years, but are still ongoing. But then I am not 13 years old. My improvement was remarked upon by a MD who works with autistic children and who is board certified in neurodevelopment.

    Reading the article in New Scientist (your link doesn’t work), this treatment was not done by a quack. It was done by a senior clinician/researcher at a well respected institution who followed all the FDA rules.

    It was also not based on a faulty chain of assumptions.

    I agree with you that inflammation is complicated. It is much more complicated than I think you appreciate. Corticosteroids are the “blunderbuss” of immune system modulation. They don’t simply “modulate” the immune system, they turn it off.

    Autism is a condition of neuroinflammation. That neuroinflammation is subtle and complex and does not respond to blunderbuss-type treatments. It also can’t be looked at easily because it is internal to the skull and is very subtle. My hypothesis is that the neuroinflammation of autism is a “feature”, a feature to try and protect the CNS from high levels of NO, such as might occur during sepsis, where high NO levels can kill any tissue with active mitochondria. Nervous tissue always contains active mitochondria, so it must always be protected during sepsis by neuroinflammation.

    The reason this man was able to do this trial on his son, was because he was wealthy. Not being wealthy, my research is proceeding very slowly.

    The mechanism by which the worms worked is by increasing nitric oxide via induction of iNOS. This is exactly analogous to the positive effects of fever therapy on multiple neuropsychiactric disorders. Fevers do temporarily resolve the symptoms of autism.

    The effects of these worms on autism is not analogous to the effects of rabies or toxoplasmosis. Rabies and toxoplasmosis cause local inflammation in the brain. Worms that stay in the gut do not.

    I appreciate your concern that quacks will jump on this and do dangerous and harmful things to children. Quacks will do that anyway.

    I don’t see any evidence that this idea has in any way gone awry.

  4. How long to they propose continuing the infestation? Am I understanding correctly–the worms are living in the host? Is there an optimal threshold of worm population vs. adverse impact on host health? If so, how do you control the worm population without killing them all off?

    Parasites are fascinating… not in my own body, but from a distance.

    • That is why I like my bacteria better. They form a self-sustaining biofilm that requires no maintenance.

      • Vicki

         /  February 11, 2011

        Unless you’ve got your bacteria photosynthesizing somehow, they are drawing sustenance from your body, as are the worms.

        This isn’t necessarily a problem: I’ve got a nice Escherichia coli colony going about its business in my gut, as do most of us. But the difference is one of scale: that your bacteria are self-sustaining in the sense of not requiring you to add more or hand them some kind of supplements could become a problem rather than an advantage.

      • My bacteria survive by oxidizing ammonia to nitrite. That is their sole source of ATP. They are incapable of growing on any bacteriological media used to isolate pathogens.

        They completely lack the enzymes necessary to derive ATP from oxidation of anything organic. They are also very slow growing, with an optimum doubling time of ~10 hours compared to 20 minutes for E. coli. If they did cause an “infection”, it would proceed with a characteristic time of months instead of days. They are obligate aerobes, so they can’t live in the anoxic gut.

        When heterotrophic bacteria oxidize amino acids, they release ammonia. My bacteria take that ammonia and oxidize to NO and nitrite. That NO/NOx is toxic to heterotrophic bacteria and inhibits their quorum sensing so they don’t express virulence factors.

        These bacteria are extremely common in the soil, and even in municipal water supplies. When chloramine is used to kill heterotrophic bacteria, it decomposes releasing ammonia. Ammonia oxidizing bacteria oxidize that ammonia. They are commensal with many organisms. Aquariums have to have them to remove metabolic ammonia or the fish will die from ammonia toxicity. There has been no recorded instance of “infection” by any of these bacteria in any organism. It is very likely that an “infection” is not possible because these bacteria lack all virulence factors. They have no pathways to produce toxins, or proteases, no transporters to excrete them, no transporters to import degraded animal tissues, no metabolic pathways to utilize degraded animal tissues except for ammonia.

        They are susceptible to antibiotics and with such a long doubling time, it would take a long time to evolve resistance. Since they don’t have the metabolic hardware to metabolize or excrete xenobiotics, the don’t have the gene diversity to elaborate on to evolve resistance to antibiotics. Even if they were resistant, they don’t cause tissue damage.

        I don’t think there is a pathway by which they can cause harm. No one has yet suggested a credible harm pathway.

      • The brown gunk that gets flushed out of municipal water pipes is usually very high in ammonia oxidizing bacteria. If the water system uses chloramine, it is almost certainly those bacteria.

  5. becca

     /  February 10, 2011

    If the inflammatory basis is really so pronounced, it seems to me that it ought to be possible to look at retinas for evidence of neuroinflammation to help diagnose autism. Of course, I would not place good money on the bet that there is only one underlying cause to autism. On the other hand, the recent increase in it does nicely mirror the recent increase in asthma and allergies- have there been any clinical trials that look at correlation there?

    There are actually a couple of studies looking at TLR responses in vitro from PBMCs from people with autism- it looks reasonable. I think there’s an altered immune response.
    Which is not the same as the immune system mediating the pathology. For all I know, normal social people are always getting extra staph from touching each other and on average autistic people are cleaner.

  6. Dear PalMD,

    Thanks so much for writing this. Even the evidence-based autism parents have been swayed by the seemingly-sound “science”.

  7. Liz, there has been no evidence presented that using gut worms like this is not based in science. There has been no evidence presented that anything has gone awry, no evidence that there have been any adverse side effects. The only reason I can understand for objecting to this is the “icky factor”. Just about every medical procedure feels icky before the physiological rationale behind it is understood.

    I am disappointed that PalMD has allowed his feelings of ickyness to cloud his analysis of this.

    Fever is known to temporarily resolve the behavioral symptoms of autism. I discuss that research in great detail.


    This is “the same kind” of thing.

    As a researcher who is working in the minutia details of autism physiology, I can tell you that there are excellent physiological rationales for this. Yes, there is an ickyness factor. That is much easier to deal with than the xenophobic feelings that autistic behaviors invoke in some people.

  8. The admittedly anecdotal cases of Crohn’s disease mitigation via parasites (the case I remember involved Cambodian hookworms) are impressive. They have led to interesting testable hypotheses regarding inflammatory pathways that may be turned on and off by parasites so that some day we can do the same pharmacologically (I personally would be a bit freaked about the worm infestation thing, but I am a bit of a princess).
    The leap to giving kids worms to treat autism is a loooong one. Crohn’s is clearly an inflammatory disease, and antiinflammatory treatments can treat it. Do glucocorticoids or other antiinflammatory drugs do anything in autism? Not that I’m aware of, although I admit this is not my area of expertise.
    Ultimately, I believe parasitic worm experiments are going to lead to new understandings of inflammation that will help treat a wide variety of disorders. Could one of these be autism? I won’t rule it out, but I’m not ready to treat a whole bunch of kids with worms, either.

    • becca

       /  February 11, 2011

      well there was that one kid with autoimmune lymphoproliferative syndrome … (PMID: 10802504)

      It’s also making me reconsider that wacky story about acetominophen being protective against stress from social rejection. I assumed if it was a real effect it was the anaglesic thing but…

    • THIS, exactly. I agree that the inflammatory Th1/Th2 stuff is totally cool and maybe one day we will be giving each other doses of parasites to regulate our immune systems.

      But I wouldn’t give it to my kid tomorrow, not without clinical trials.

      • becca

         /  February 17, 2011

        I’d worry more about reversibility of the effect than it standing up to clinical trials. For one thing, diagnosis of autism being what it is, you could easily end up with a good treatment for a *subset* of autism not showing any impact in a broad-population clinical trial.
        If you can tilt the immune response *back* when needed (there are some very good reasons your body can do both Th1 and Th2), and cure the worms, or use ones that don’t survive long in humans, then there’s minimal harm involved.
        The ickiness and unproven nature don’t bother me. It’s the potential to skew this stuff without realizing what we’re doing and render people more at risk for other diseases that worries me. Importantly, that kind of thing could happen even if it passes clinical trials- it might be a very long-term effect. Kind of like my parents caring more about how *old* a particular pharmaceutical intervention was rather than the current data on efficacy. If it works, great, but you wanna be sure it doesn’t give you cancer and it takes huge populations with many exposure years to get the best data.

        I probably wouldn’t give my kids worms. But if I was the one with the diagnosis, I might consider it.

  9. Dr. Dredd

     /  February 11, 2011

    FYI: The link to the article in The Scientist doesn’t work. 🙂

  10. Jess

     /  February 11, 2011

    @Anthropologist: While I know very little about the biology of autism and even less about worm treatment for it, I’m at least somewhat familiar with the way experimental TSO treatment is handled in MS patients. The worms do live in the host – the patient drinks about 2,500 worm ova (not larvae as suggested in the post) in a small amount of buffered solution once every two weeks. The worms hatch in the host and grow a bit, but since these are pig-specific worms, they don’t thrive, their numbers aren’t maintained over time, and they don’t reproduce. That makes the number of worms in the person easily controllable by how many ova are administered. Since Trichuris suis doesn’t grow well in the human host (vs. Trichuris trichiura, T. suis’ human-specific cousin), there’s nothing that the patient produces/sees that looks… out of the ordinary.

  11. @ Jess: really interesting! Thanks!

  12. Doc Frankie

     /  February 12, 2011

    I can understand where PalMD comes from. One can’t be part of a blog which mission is to body slam, trounce and make mince meat of denialism, quackery and all sorts of fuckology so prevalent nowadays without seeing any deviation from normal channels of clinical research with a high index of suspicion.

    That said, this idea “gone horribly awry” is yet another example of thinking outside the box in the face of quiet despair brought by YEARS of unsuccessful treatment of a condition that can easily destroy a person AND his/her family. Just try to live with a kid afflicted with autism and come back to tell me how long you’d be willing to wait for the conventional scientific thinkers to come around and do something “different” while you’re the one stuck with all the problems.

    There are other examples like this one: Lorenzo Oil comes to mind. AIDS victims got sick (literally!) and tired of waiting for all these lace-pantsy clinicians who were more preoccupied by protocols and procedures than anything (and anybody) else.

    Bottom line: Faced with a seemingly intractable problem, science does not resolve them by forming committees, but by quantum leaping the conventional knowledge and accepted ideas to land unto the unknown and take it from there.

    But for that, one needs to have the courage to fail.

  13. Solitha

     /  February 13, 2011

    I read both stories, thanks to Ed Yong.

    All I have to say to this one is, wow… did your logic fail in a majorly hypocritical way.

    Firstly, you’ve misrepresented syllogism points 2 and 3. Autism as a whole was not the intended target, but the violent, disruptive behavior that seems to be related.

    Secondly, while points 1 and what should have been 2 (that the behavior was inflammatory-related) may not have been proven *true*, they certainly haven’t been proven *false*.

    This guy didn’t just grab his son and fly off to Europe to avoid all the process like some would do. He did his research. He made a hypothesis. He presented the work to his son’s doctor. They worked with the FDA as well as the hospital. Every step of the way, the work was being checked and checked again.

    Actually, it sounds like he went through a lot more checking than this article did.

%d bloggers like this: