Antonio
Hardan, the psychiatrist at the Stanford School of Medicine, has published
another paper. Hardan is interesting, he
is a clinician rather than a rocket scientist, but he gets involved in a very
wide variety of clinical trials, usually of existing drugs that might be
effective in autism.
In his latest paper, this time about Glutamatergic Dysfunction in
Autism, he highlights the problems with clinical trials:-
·
Heterogeneity
of autism
·
Subjective
rating scales rather than biological measures.
In other
words there is no single autism and there is no good way to reliably measure
the efficacy of any drug tested on it. Consider what that really means.
Hardan really should know about this, just look at the clinical trials he has been
involved in:-
So why bother with Clinical Trials?
This may
sound like a very unscientific question, but perhaps it is not. A couple of years ago Roche pulled the plug
on Arbaclofen, because it “failed” in its autism clinical trial. Many parents thought it worked. Now the Simons Foundation has acquired the
rights to the drug and is restarting trials.
How many other trial drugs were prematurely brushed aside?
Many years
ago the hormone secretin was put forward as a therapy for autism, particularly
for people with GI problems. Several expensive clinical trials later, it was determined to be ineffective. But some people continued to rave about it. Where they all deluded?
The very
expensive IVIG therapy has also been put forward as a wonder therapy for
autism. The critics highlight that in
studies 90% of people do not benefit and therefore the therapy has little
value. But what if you are in 10% that
do respond very well?
Abstract
Since autism
has been associated with immunologic abnormalities suggesting an autoimmune
cause of autistic symptoms in a subset of patients, this study was undertaken
to investigate whether intravenous immunoglobulin (i.v.Ig) would improve
autistic symptoms. Ten autistic children with immunologic abnormalities,
demonstrated on blood tests, were enrolled in this study. Their ages ranged
from 4 to 17 years, with two girls and eight boys. Eight children (1 female and
7 male) historically had undergone autistic regression. Intravenous
immunoglobulin, 200 to 400 mg/kg, was administered every 6 weeks for an
intended treatment program of four infusions. In five children, there was no
detectable change in behavior during the treatment program. In four children,
there was a mild improvement noted in attention span and hyperactivity. In none
of these children did the parents feel that the improvement was sufficient to
warrant further continuation of the infusions beyond the termination of the
program. Only in one
child was there a very significant improvement, with almost total amelioration
of autistic symptoms over the time period of the four infusions. Once the treatment
program was completed, this child gradually deteriorated over a 5-month time
period and fully reverted to his previous autistic state. In this treatment
program, five children had no response to intravenous immunoglobulin. In the
four children who showed mild improvements, those improvements may simply have
been due to nonspecific effects of physician intervention and parental
expectation (ie, placebo effect). However, in one child there was a very
significant amelioration of autistic symptoms. There were no distinguishing
historic or laboratory features in this child who improved. Given a positive response
rate of only 10% in this study, along with the
high economic costs of the immunologic evaluations and the intravenous
immunoglobulin treatments, the use of intravenous immunoglobulin to treat
autistic children should be undertaken only with great caution, and only under
formal research protocols.
Just in this
blog, which is amateur and not intended as a rigorous scientific review, we
have seen numerous “rare” conditions that lead to “autism” that are actually treatable.
If you add
up all these “rare” conditions you get a sizeable proportion of all the autism,
diagnosed in those under four years old (i.e. more severe autism).
Clinical Investigations
If you
accept that the initial autism diagnosis really tells very little, then you are
left, like Hardan, testing all sorts of clever ideas on a trial group of kids
who may have one to several, of thousands of discrete dysfunctions (CNVs etc.).
Then if you
get a 10% response rate, you are doing great.
If you
target something like oxidative stress, that is caused by hundreds of those
thousands of discrete dysfunctions (CNVs etc.), then your odds of success shoot
up. This was the case in Hardan’s trial
of N-acetyl cysteine.
Hardan is
now going to trial oxytocin on kids with autism, but this idea has already been
well and truly “trashed” by highly respected mainstream doctors. They do this because they think autism is
something easy to define and measure like high blood pressure. If it is therapeutic in 10% of cases, that is
great.
Quacks, Off-label and Clinical Investigations
I think it
is great that Hardan can try all these drugs at Stanford and nobody even thinks
of calling him a quack. The same applies
to a small number of inquisitive doctors at Johns Hopkins and Boston Children’s
Hospital.
It would be interesting to know how Hardan treats his patients with ASD, who are not enrolled in a clinical trial. Does he prescribe off-label?
It is clear
that most doctors in developed countries will run a mile/kilometer at the idea
of treating somebody off label. They
fear being struck off/sued/ridiculed.
We had the UK pediatrician commenting on this blog that
Baclofen was effective in 70+% of her/his patients with anxiety plus
Asperger’s, but did not feel happy to continue prescribing it without some
supporting evidence from elsewhere. The
fact that it was safe and effective was not enough.
Many of the
tiny number of off-label doctors really do look like quacks to me, so I can
understand the concern of mainstream doctors not to want to be associated with
them.
What is the,
scientifically well-briefed, parent supposed to do? (if self-treating is not an
option)
I think
there should be a way where you can enroll your child in a “clinical
investigation”, where you accept that all the treatments are experimental and
therefore have a higher level of risk than normal. You waive your right to sue the doctor, or
the hospital. You can opt out of up to
10% of the therapies, based on valid concern.
For example, you might think IVIG is not safe.
You then
enter a program in which all your child’s data can be used for research
purposes. So you agree to have to have
EEGs, scans, genetic testing, spinal tap/lumbar puncture, blood tests, urine
tests, hair tests etc.
The child is
completely profiled and material is stored for possible further analysis later.
All known tests are then carried out, even obscure
things like biotin deficiency, creatine deficiency and those amino acids we saw
that triggered rare autism.
Then you go
through all of the therapies known to be effective in some people. So it includes memantine, IVIG, donepezil, bumetanide, oxytocin, propranolol,
baclofen, arbaclofen, even Zyrtec, NAC, D-Cycloserine, carnosine, carnitine,
pancreatic enzymes, probiotic bacteria etc.
The whole
process would take a year. If you
treated 1,000 children you would then have a wealth of data.
You might
have individually rare disorders totaling 15% of cases and then several
clusters where the same drugs were effective in sizeable groups of
children. Then you would be able to look
back in the data for the biomarkers of each cluster.
Then you
would write a smartphone app for doctors to treat autism. They would input the various biomarkers requested
and out would come the suggested drug therapy recommendation(s). So it would be a “guided off-label” approach
where the doctor knows that the recommendations are “scientifically supported”
but may not be perfect.
We just need
the Simons Foundation to sponsor it!
If you think it might be too expensive, just remember that at the recent international autism conference in Utah, there were 2,000 scientists and researchers in attendance. What exactly have they achieved, in practical terms, in the last 10 years and are likely to achieve in the next 10 years?
It does seem that some view success as diagnosing ever more people with "autism", so that they can receive "services", when they really should be diagnosing specific biological dysfunctions.
It is not an easy task, but you do not need 2,000 researchers. You just need 20 pragmatic people to review the data and make a decision tree showing how to choose the 5 drugs most likely to help a particular person, based on their specific biomarkers.
I guess that would leave 1,980 people with not much to do.