The NIMH has abandoned its plans to perform a study of chelation therapy in autism. In my opinion, this was the right move. It also appears that the decision was influenced by widespread criticism from many quarters, including the science blogging community, that such a study would be unethical.

I wrote about the NIMH plans to study chelation in autism for Science-Based Medicine.  In that article I brought up the main points of criticism – that the risks of chelation therapy were not outweighed by the probability of it being effective for autism, an unfavorable risk vs benefit ratio means that studying it in a vulnerable population (autistic children) entails special ethical problems, and that the primary stated goal of the study – to counter anecdotal reports of the treatment’s utility – would not be accomplished. The anti-vaccine mercury militia is not a group persuaded by scientific evidence, and so performing a scientifically and ethically dubious study to satisfy them is a fool’s errand.

Apparently, the NIHM got the message.

According to reports:

“There will be parents who are disappointed,” said Richard Nakamura, the scientific director of NIMH. “We recognize that for children there is a fine line for the risk-benefit ratio. You have to be pretty certain of the overall safety of the procedure.”

The reaction of the anti-vaccine crowd – those who believe that mercury in thimerosal in some vaccines (now removed) caused some cases of autism, is predictable. Rebecca Estepp, national manager of Talk About Curing Autism, a support group for families with autistic children, said:

“By discontinuing this study, the NIMH will not prove the effectiveness of chelation therapy one way or another. Instead, they have merely left parents with more unanswered questions.”

However, history clearly indicates that doing such studies does not end the controversy, therefore there is no benefit to performing the study to justify the risks to the subjects. The study can only be justified if there is a firm scientific basis for the possibility that chelation might work for autism. But the treatment is based upon the failed hypothesis that autism is mercury toxicity. Without this premise, there is no way to justify the study.

The anti-vaccine propaganda site, Age of Autism, also had a predictable response:

So who canned the NIMH chelation study as “too dangerous?” Children are given huge doses of chemotherapy and radiation in a desperate effort to save them from cancer – fully knowing the side effects themselves can be deadly.  It’s a fair risk most parents are willing to take to help a sick child.

Typically, they miss the point that medical interventions are judged based upon risk vs benefit. Chemotherapy with known statistical benefit can be justified to treat a cancer that, if untreated, will have a horrible outcome.

This calculation does not apply to chelation therapy for autism, because there is no basis upon which to base any anticipation for benefit. Therefore the risks are not justified – even experimentally (which has a lower threshold than recomminding a treatment for widespread use).

Proponents, of course, site the anecdotal experience of parents treating their autistic children with chelation therapy as sufficient justification. However, such evidence is so weak as to be almost worthless. Perhaps this would be sufficient to perform a study of a benign treatment, but not one as risky as chelation therapy. Anecdotal evidence is extremely unreliable because it is uncontrolled and subject to a variety of biases. Autism is especially vulnerable to generating the false conclusion that a treatment was effective, because children with autism can improve normally as they age and mature. This background maturation would need to be separated from any treatment effect – and this simply cannot be done anecdotally. There are also other psychological factors, such as expense justification – parents need to justify to themselves the expense and risk of putting their child through chelation therapy, so they convince themselves that it helped.

The only way to know if a treatment actually works is through controlled clinical trials. Then why not perform a study of chelation therapy? I would certainly like to have the results of a well-designed and executed study of chelation in autism. But we have to consider the expense of getting that information compared to the probable utility. The NIMH decided they could not justify the risks of the treatment, and the money for the trial could be more productively spent on other studies. This hits upon the two key factors in deciding what to study. The first, as I stated above, is risk vs benefit.

The second is the allocation of finite resources.  These resources include research dollars, but also researcher time, and the availability of subjects to study. Government agencies tasked to spend taxpayer money efficiently in deciding what research projects to undertake have to consider the utility of a study compared to all other studies seeking funding. The science and ethics simply did not support funding this study over others.

I am glad to see that the right decision was made in this case.

Note: David Gorski also reported this development at Science-Based Medicine.