Adam's Recent Study on Mercury in Baby Hair of Autistic/Control Children

The question that has been driving the "mercury-autism" research question for the past decade is whether those suffering from autism are carrying a higher body burden of mercury than "normal" children and therefore, whether autism could be properly characterized as an inability or a compromised ability to excrete mercury. Bernard's 2001 study suggested that the signs of autistic behavior and mercury poisoning corresponded closely, but other studies disputed that. As a way of trying to break the impasse, in 2003 Dr. Amy Holmes examined the mercury content in first cuts of baby hair in a number of autistic and neurotypical children and contended that neurotypical children had an eightfold higher level of mercury in the hair than did autistic children. She also argued that the severity of a child's autistic symptoms correlated with a lower mercury level in the hair. She concluded there was also a statistically significant difference between the levels of fish consumption or number of dental amalgams of the mothers of each group. Thus, it appeared that her study gave fuel to and supported the 2001 Bernard article. Since failure to excrete mercury was strongly correlated, in her mind, with autism, the question that her work naturally provoked is whether excessive mercury in the body causes autistic symptoms.

Holmes' work received a fairly serious attack from Dr. Brent in the Omnibus Autism Proceeding, in his June 22, 2007 testimony. His contention was that her numbers were hugely skewed and that no one had remotely been able to replicate them. Perhaps fearing that this kind of attack was in the offing, Dr. James Adams of Arizona State Univ., working with a team of others, tried to replicate Dr. Holmes' Study. His results of first-cut baby hair study were recently published as "Mercury in first-cut baby hair of children with autism verses typically-developing children," Toxocological & Environmental Chemistry (2008), 1-14.

Adam's Conclusions in a Nutshell

I have not yet met Dr. Adams, but all I have seen of his work and his style of operation is that he is a careful and diligent researcher, one who is loathe to "overread" data based on a preconceived theory. What he does in this article is to give us numbers which challenge Holmes' conclusions pretty decisively, but he frames the conclusion in the article's abstract in a way that tends to muddy the waters. That is, his statistics show that the hair of normal/neurotypical babies did not have a significantly higher level of mercury than that of autistic children. Yet, he couches his conclusion, in the abstract, in words that might tend to mislead a hasty reader. Here are his words:

"Logical regression analysis showed that compared to children with higher levels of mercury (above 0.55 mcg/g), children with lower levels of mercury in their hair (below 0.55 mcg/g) were 2.5 fold more likely to manifest with autism."

But that isn't really the point. The point is not relative levels of autism as it relates to an arbitrary concentration of mercury in the hair, but a comparison between the average level of mercury in the hair of the two groups. Indeed, the appearance of "2.5" might make someone immediately compare it with Holmes' "8" and think, "Hm, not exactly the same, but still showing a huge body burden of mercury in autistic kids," but that comparison would be inappropriate. I know of no other study that picks an arbitrary number (0.55 mcg/g) and talks about manifestations of autism below and above the line. The focus is to compare mercury in the hair of these two groups of babies. Well, what does Dr. Adams' data show when he does that?

The Mercury Hair Burden

After going through his impressive methodology, which included sending the hair samples through the same process used in the NHANES study cited in this essay, and sending duplicate samples (when there was enough hair) to another lab and correlating the findings (things correlatd), Dr. Adams discusses the levels of mercury in baby hair. He compared the numbers in his study to those of Holmes' study:

"The average level of Hg in the baby hair of children with autism is similar between the two studies (...0.73 mcg/g in this study [excluding the four outliers] verses 0.47 mcg/g in the Holmes et al. (2003) study. However, the level of Hg in the baby hair of typically-developing children of this study was much lower than that in the Holmes et al. (2003) study (...0.87 mcg/g in this study verses 3.63 mcg/g in the Holmes et al..."

Let's pause for a moment. Adams is suggesting that the ratio of mercury in autistics hair/mercury in 'normal' baby hair in his study was 0.73/0.87, while in the Holmes study it was 0.47/3.63. Thus, if we compared apples and apples, so to speak, we conclude that the ratio is 1:1.2 compared with 1:8. The difference is stark. In fact, the numbers make it look as if the Holmes study was almost completely refuted. That, at least, is how I read Dr. Adams' numbers.

Looking More Precisely

One of the conclusions of the Holmes study was that the amount of mercury in baby hair correlated with severity of autistic symptoms; i.e., the more severe the symptoms, the lower the amount of mercury in the hair. Adams et al. said they could not replicate this finding. In addition, Holmes had argued that the mothers of autistic children had higher levels of fish consumption and dental amalgams than "control" mothers. Adams, however, found that these differences were not significant. The only significant difference Adams found between the autistic and neurotypical infants was that the autistics had been exposed to a significantly higher level of childhood antibiotics than the controls. But this factor wasn't even mentioned in the Holmes study. Thus, we are in the curious position of concluding, with the Adams study, that almost all the things that were "red flags" for Holmes are not so for Adams. Thus we can understand the rather weak language in his article, summarizing earlier scholarship:

"Overall there is a substantial body of evidence that suggests Hg poisoning may be involved in some cases of autism..."

And his conclusion:

"early exposure to Hg appears to be involved in the etiology of autism."

But his data don't sound the clear trumpet signal. Or really even a trumpet signal at all. The most that they say is that there is still a little less mercury burden in the hair of autistic babies than of neurotypical babies, but this may be due, he suggests to ingestion of antibiotics. But antibiotics, as he says, have more to do with gut flora than mercury production.

Conclusion

So, where does that leave us? Well, though I could say that the mercury-autism link has not been proven, I might better say that it has now nearly been disproven. Adams' data really starkly contrasts with that of Holmes, especially on the issue of mercury level in the hair of neurotypical babies. The major reason for this difference is probably the careful methodology used by Adams. His identification of higher levels of antibody ingestion by autistic babies than neurotypical babies suggests a frutiful avenue of future research. But unless someone else wants to take on the mercury in hair issue or mercury burden in the body issue to try to demonstrate that autistic kids excrete a much lower level (and therefore retain a higher level), I think the issue is almost moot now, even though Dr. Adams seemingly wants to keep it alive by the way he has framed his conclusions.

3760