Autism Street

Remember this great poster from
Fox Mulder’s FBI office in the
TV series the X-Files?

Sometimes I think I should try
“believing” rather than looking
at available evidence, but I just
can’t do it. There always seems
to be some facts that get in the way.

A ‘new’ autism hair study about
toxic metals and essential minerals
was recently published and indexed
on Pubmed.

I say ‘new’, because this doesn’t appear to be a ‘new’ study at all. It’s been available on Jim Adams’s website at ASU for some time, as it was apparently originally written three years ago. Excerpted from the (now published) paper’s abstract:

The objective of this study was to assess the levels of 39 toxic metals and essential minerals in hair samples of children with autism spectrum disorders and their mothers compared to controls.

Iodine levels were 45% lower in the children with autism (p=0.005). Autistic children with pica had a 38% lower level of chromium (p=0.002). Autistic children with low muscle tone had very low levels of potassium (–66%, p=0.01) and high zinc (31%, p=0.01). The mothers of young children with autism had especially low levels of lithium (56% lower, p=0.005), and the young children (ages 3–6 yr) with autism also had low lithium (–30%, p=0.04).

Further investigations of iodine, lithium, and other elements are warranted.

“So what?”, right? Well maybe “So what?”, however, some may find portions of the discussion section of the paper interesting. Here’s an example:

Overall, it appears that the children with autism do not have major differences in their levels of toxic metals compared to controls.

Yet in the very same section it concludes:

Thus, our results are not necessarily inconsistent with the results of Holmes et al.

So Holmes et al. says that some autistic children have way lower mercury in their hair than controls, and Adams et al. says that major differences weren’t found, but this is “not necessarily inconsistent with the results of Holmes et al.”? Huh? Let’s look a little more of what Adams et al. have to say about the Holmes et al. study.

In addition, there was a recent study (11) of the level of mercury in the baby hair of infants (aged 12–24 mo) who later were diagnosed with autism compared to controls (n=94 and 45, respectively). This study found that the autism group had one-eighth of the normal amount of mercury in their baby hair compared to controls.

One eighth of the “normal” amount of mercury?

Well, not exactly. It depends what is meant by “normal”. The Holmes et al. study found that the autism group had 0.47 μg/g (+/- 0.28 μg/g), which is approximately one eighth of the levels in the Holmes et al. control group. But, was the Holmes et al. control group normal? Let’s compare these same reported findings to a very large control group consisting of data from a study published in 2004 as part of the 1999–2000 NHANES study of 838 children aged 1–5 yr. The reported geometric mean was .12 μg/g (+/- .02 μg/g) and the reported arithmetic mean was .22 μg/g (+/- .04 μg/g), making the Holmes et al. results quite similar to, if not higher than, the “normal” amount of mercury - not one eighth of it.

The interesting thing here is that the Holmes Baby Haicut Study reported a control group (45) average of 3.63 μg/g (+/- 3.56 μg/g). That looks like over fifteen times the very large (838) group arithmetic average of .22 μg/g and thirty times the geometric mean of .12 μg/g from the NHANES study. Is this an anomaly?

Adams et al. continue with:

In the control group, the amount of mercury correlated with the mother’s seafood consumption and number of mercury amalgam dental fillings, but that was not true of the children with autism, who had a low level regardless of their mother’s seafood consumption or number of dental fillings, which suggests a general inability to excrete mercury.

Except that they apparently didn’t really have a “low level”. In fact, they had a level quite similar to, or higher than the NHANES average, and higher than the average of both typical and children with autism published in Adams et al. This probably suggests that the “inability to excrete mercury” hypothesis could possibly, if not more likely, be post hoc correction as suggested by blogger Prometheus.

And this:

They also found that the severity of autism had a strong inverse relationship with the level of mercury, with the most severe group having the lowest levels of mercury in their hair. This is consistent with the hypothesis that the group with the most inhibition of mercury excretion would be the most severely affected.

Except that they apparently didn’t really have “low levels” of mercury in their hair. As far as excretion goes, “Hair incorporates Hg present in circulating blood during hair formation in the hair follicle.” Source.  It may not necessarily be an indicator of mercury “excretion” at all.

Adams et al. on the validity of their study:

In terms of the validity of our testing, it should be pointed out that the mean values we found for the typical children aged 3–6 yr (0.21 µg/g) are similar but somewhat higher than those of the 1999–2000 NHANES study (4) of 838 children aged 1–5 yr (0.12 µg/g). In terms of our results, our finding of similar values of mercury for autistic children and controls is consistent with the study by Ip et al. (10), which found similar (albeit much higher) levels in the autism and control children in Hong Kong.

It should be pointed out that mean values found by Adams et al. for typical children are nearly identical to the findings of the 1999–2000 NHANES study of 838 children - Adams et al. (.21 µg/g) vs. McDowell et al. arithmetic mean (.22 μg/g). This does in fact suggest that the Adams et al. data is valid. It should also be pointed out that both Holmes et al. and Adams et al. apparently had hair samples analyzed at Doctor’s Data. It’s possible that both studies were even conducted around the same time (somewhere near 2002) presumably, or at least possibly with the same epuipment and methods too, yet they both got vastly different control group results.

Adams et al. conclude this section with:

Thus, our results are not necessarily inconsistent with the results of Holmes et al. (11), which found unusually low levels in baby hair, as the ages of their group (12–24 mo) are quite different than ours (age 3–15 yr). Actually, if both sets of data are valid, then they suggest a temporary loss of the ability to excrete mercury in young infants. This temporary loss could be explained by higher use of oral antibiotics (for ear infections) in children with autism as we found here, as Rowland et al. (15) showed that oral antibiotics dramatically inhibit mercury excretion to one-tenth of normal in rats.

Except it’s pretty clear that Holmes et al. probably didn’t find unusually “low levels” in baby hair. It would seem that what they found was an unusual control group.

This would suggest that both sets of data may not be valid, and that the suggested “temporary loss of the ability to excrete mercury” may be more along the lines of some sort of new twist on undead bad science, but I could be wrong. I couldn’t find data about how many kids in the ear infection subgroup were treated with oral antibiotics vs. topical antibiotics if treated with antibiotics at all (subgroupings were apparently made based on parent report only), but I did find this in the Adams et al. paper

A recent study (10) measured mercury levels in the blood and hair of children with autism (n=82; mean age =7.2 yr) compared to controls (n=55) living in Hong Kong. The authors found that children with autism had slightly elevated levels of mercury in their blood, but the difference was not statistically significant (p=0.15). They did not find a significant difference in hair mercury levels, but the levels were unusually high in both populations (2.26 and 2.07 ppm for the autism and control groups, respectively); the authors suggested that the reason for the high levels is that the “Hong Kong Chinese are famous for eating seafood.”

[Emphasis mine.]

If 2.07 ppm is “unusually high”, what does that suggest about the Holmes et al. control group’s 3.63 ppm? Unbelievably high?

Note: No implication that the real ”Gee Your Hair Smells Terrific” shampoo contains or ever contained mercury is intended. The point of the farcical image is to highlight the noticeability of the control group in the Holmes et al. study. From what I understand, “Gee Your Hair Smells Terrific” has not been manufactured by Jergens (a division of KAO Brands) in the U.S. for years. If you’re into to terrific-smelling hair, apparently you can still buy the real “Gee Your Hair Smells Terrific” shampoos and conditioners in the Philippines, on eBay, or from some select importers.