Here is part of an interview with Mark Blaxill, research chair of the group SafeMinds, which advocates removal of mercury from medical products.
Blaxill, who lives in Cambridge, Mass., is a business consultant who began doing his own research on autism after his daughter, Michaela, 9, was diagnosed with the disorder.
Blaxill said SafeMinds' analysis of government data shows children who were exposed to a mercury-based preservative in vaccines had a much higher rate of autism than those with zero exposure. Subsequently, the Centers for Disease Control and Prevention in Atlanta put the study through several more generations of data inclusions, exclusions and analysis; the final study did not find an increased risk for autism.
The CDC found that the early data and analysis cannot be used to draw conclusions, and the agency stands by its study -- a view endorsed last year by the prestigious Institute of Medicine of the National Academy of Sciences.
Still, the issue is increasingly controversial -- environmental lawyer Robert F. Kennedy Jr. recently called the CDC's studies "cigarette science," referring to flawed and self-serving reports from cigarette manufacturers on the health risks of tobacco.
Most medical experts and government health officials say repeated studies rule out thimerosal, the ethyl mercury preservative, as a culprit in the rise of autism diagnoses in the 1990s. Thimerosal was phased out of childhood vaccines in the United States in 1999, though it is used widely around the world.
Blaxill called the material that SafeMinds received from the CDC "Generation Zero."
In the interview, he referred to the CDC analyst who led the study, Dr. Thomas Verstraeten, and to e-mails Verstraeten wrote to other CDC researchers.
SafeMinds did a whole series of Freedom of Information requests on the CDC's work on thimerosal, and we got a pile of printouts in the raw spreadsheet form -- the output that comes off the statistical program when you're working numbers.
What was interesting is that if you really dug into the numbers, the ones that showed up as statistically significant all happened to be -- if you deciphered the code -- the high-exposure population group (to thimerosal-containing vaccinations).
When you do that, and also begin tracing the e-mail traffic around the time these documents and spreadsheets were being sent around the CDC, it demonstrated something that I found pretty stunning. The first time Thomas Verstraeten sat down to look at thimerosal exposure -- unencumbered by study protocols and exclusions and a whole set of convenient "screens" that you could put on the data -- he found very, very profound effects from mercury exposure. And it was not just autism, but a whole range of developmental disorders.
It led him to obviously be quite concerned, from the e-mails he was sending around. From that day, he was the guy who was most concerned in all the CDC group, and I think you could probably trace that to his original exploration.
In November of 1999, right around Thanksgiving, he did a long list of possible outcomes and also had a fairly broad set of exposure dimensions that he used. There was one really useful category of exposure, which went all the way from zero thimerosal to relatively high exposure.
At the very highest level of exposure, almost every one of the neurodevelopmental disorders -- autism, ADD, ADHD, speech delay, sleep disorders, tics, stammering -- had elevated relative risks on the order of 4 to 8 times, and the vast majority of those risks were statistically significant.
So you had a population that was obviously damaged along a wide range of outcomes at the highest level (of mercury exposure).
That was a strong signal in November of 1999. Then he went back in December and pruned the database down a little bit and looked at a shorter list of variables. He must have made some changes in the database, and this time he emerged with the highest exposure category for autism showing an 11 times increased risk -- and no increased risk for the control group.
That was pretty stunning. An 11-fold relative risk for the highest thimerosal exposure is roughly equivalent to the tenfold rate of increase in autism in the 1990s.
He was obviously disturbed. He didn't want to overreact. I would have the same kind of reaction he had, which was not wanting to get out on a limb but being concerned.
As the parent of an autistic child looking inside the "sausage factory," what I see is a little analyst sitting at a computer, and lo and behold, he runs the numbers unencumbered by any quote-unquote improvement, and he finds something that is stark and concerning and strong.
And then by the time the final publication comes out four years later, they have reduced that signal to almost statistical insignificance.
And then you read the conversations they're having along the way -- you read the e-mails, the transcripts of meetings -- and this is not a group that is saying, "Oh my goodness, we're really worried about autism, we're really worried about the neurodevelopmental effects. We need to understand precisely what's going on here."
No, what you see is a group that is deeply concerned, deeply biased that this must not be true. They're looking for the flaws in the signal. You never find someone questioning an adjustment that reduces a statistical finding. All of the outcomes of the various iterations move in one direction. They almost never increase and get a stronger signal.
And frankly, that just doesn't pass the sniff test as an analyst. Almost always, when you get a signal that has some strength and is biologically plausible, that strength of signal doesn't vanish. It usually tells you something.
Watching the sausage factory, my feeling was, oh my lord, these people don't care what's going on with the kids. All they care about is covering their (backsides). I hate to say it, but it's quite discouraging to watch.
In upcoming columns we will look further at the CDC's internal data and dialog about vaccines and autism.