Autism Street

A Tale Of Two Tails

November 17, 2007 by Do'C  Printer-Friendly Version

94 Comments »

1. Comment by Schwartz — 18 November, 2007 @ 12:30 am

   Excellent Article — especially so, because you clearly delineate the issues of fact vs opinion.

   If the outliers are kept in, and a one tailed test used, does the result end up being significant, or do they even do this?

   Another question. If one repeated this type of study n times, and arrived at the same result — not enough correlation to be significant, but a similar correlation every time. Does that change anything?

   ON a side note, I’m glad you review the comments, as I find many studies are severely misleading in that department. I often wonder if the peer reviewers just gloss over that stuff?

2. Comment by Ms. Clark — 18 November, 2007 @ 12:42 am

   Thank you, Interverbal and D’oC.

   So one wonders if this paper was generated in the hopes that it could be presented in the Omnibus hearing. Maybe DeSoto or Hitlan had hopes of getting those $250 an hour expert fees or something. Maybe I’m being unkind about their motives, but they sure don’t seem to be applying any known science or statistics to what they produced in this paper. How is it the normal kids in the Holmes data set were walking around normal with massive amounts of mercury in their systems? Their hair levels were astronomical, not just high, but they were “normal” by definition. And the autistic kids had normal amounts of mercury in their hair. It really boggles the mind that this paper continues to be cited, for instance, in the Hazlehurst and Snyder cases.

3. Comment by isles — 18 November, 2007 @ 1:17 am

   Wonder if DeSoto and Hitlan are enjoying the sensation of having their hineys handed to them by a couple of bloggers.

   They acknowledge that Fido and Al-Saad (2005) found results contradictory to their claims, but don’t mention the Wecker article from 1985 finding no significant different in hair mercury.

   And another snipe - when I took statistics, indeed one-tailed tests were strongly frowned upon, unless you had a durn good reason.

   But more importantly, if hair doesn’t excrete, what the heck is this paper even about?!

4. Comment by Bartholomew Cubbins — 18 November, 2007 @ 6:13 am

   For the life of me I cannot figure out why DeSoto didn’t simply do multiple analyses (1 and 2 tail) on both the original typo-laden dataset and the corrected data and compare the results. I’m left to conclude that they’re either ignorant or had an agenda. I’ve never heard of someone submitting an article critiquing a paper without showing, using multiple strategies, that there’s a change in the outcome. Another idiot here is the editor.

5. Comment by Interverbal — 18 November, 2007 @ 9:32 am

   Hi Schwartz,

   “If the outliers are kept in, and a one tailed test used, does the result end up being significant, or do they even do this?”

   They didn’t calculate it. There were two outlier data points in this study. The bigger of the two was an unusually high mercury level for someone in the autistics group. The authors felt that by removing this, the two groups were brought closer together, but no stats were run.

   “Another question. If one repeated this type of study n times, and arrived at the same result — not enough correlation to be significant, but a similar correlation every time. Does that change anything?”

   Not in my opinion. If you can’t reject an alpha (type I error) then there is reason not to be confident in the results. However clinically speaking, there may still be something here occuring on a statistically insignificant level.

   But….. that seems to fly in the face of the arguments by DeSoto & Hitlan. They argue this issue is important because the cases of autism have risen so much. They need a big cause (or at least a couple mid size causes), one that strongly push the cases up. Having just under statistical significance doesn’t match up with what the authors need to justify their argument.

   To be logically intact, something would have to give. The advocates couldn’t say that a statistically significant portion of autism is caused by mercury and that repeated, close to (but not quite) significant outcomes prove this. They don’t match up. The advocates would have to modify their argument.

6. Comment by Schwartz — 18 November, 2007 @ 12:12 pm

   Interverbal,

   If they didn’t even bother to do an alternate run of the numbers using their suggested methodology to see what the outcome would be, then I agree with you, Do’C and a previous commenter that this study is pretty meaningless. They could have published a letter to the journal if they only meant to critique.

   The only criticism of merit appears to be the choice of the hypothesis if indeed we decide that in a case like this, two-tailed are more appropriate. At a minimum, the hypothesis should have matched the test. That’s about the only valid criticism of the original study I can see from your writeup, and like I said, a letter to the journal would easily suffice for such a criticism.

   I am surprised the initial peer review wouldn’t pick up the conflict between the choice of 2 tailed test and hypothesis though. I looked up these statistical descriptions and it matches exactly what Do’C stated, so it appears that someone did mess up there on a pretty fundamental statistical item in the original study.

   Again, excellent work by you and Do’C. I am glad you both take the time to delve into these studies (both supporting and non-supporting of your positions) in detail, as this is a practise that is sorely needed in all cases.

7. Comment by Maureen Fischer — 18 November, 2007 @ 2:50 pm

   Isn’t the Journal of Child Neurology a peer reviewed journal? Are any of your criticisms being prepared for this journal so that the authors will have a chance to defend themselves in an open forum?

8. Comment by Interverbal — 18 November, 2007 @ 4:06 pm

   At this time I don’t think we have any plans to write a letter to the editor. Speaking only for myself, I am anxious to see how Ip et al respond to DeSoto & Hitlan.

9. Comment by Do'C — 18 November, 2007 @ 4:07 pm

   Isn’t the Journal of Child Neurology a peer reviewed journal?

   I would appear that it is.
   About the Journal

   Are any of your criticisms being prepared for this journal so that the authors will have a chance to defend themselves in an open forum?

   I can’t speak for Interverbal, but I’m not preparing any criticism for this journal. The authors are more than welcome to join the discussion here if they’d like. Criticism of our opinions and ideas is welcome.

   P.S. Thanks anonlurker for the note about the typo - it’s fixed.

10. Comment by culvercitycynic — 18 November, 2007 @ 8:21 pm

    Good job, guys. Something that AD posted somewhere else, on this topic, got me to wondering whether or not any of the researchers (in any of these “hair studies”) controlled for the ratio of protein — in relation, of course, to carbs and fats — which each child consumed in their diets….or would that not even matter?

11. Comment by Do'C — 18 November, 2007 @ 9:17 pm

    Something that AD posted somewhere else, on this topic, got me to wondering whether or not any of the researchers (in any of these “hair studies”) controlled for the ratio of protein — in relation, of course, to carbs and fats — which each child consumed in their diets….or would that not even matter?

    If I remember correctly, another paper by Ip et al. in 2004, made a notation about blood and hair mercury levels correlating with frequency of seafood consumption.

    I don’t have a copy of Ip et al. (2004) from J Child Neurol handy.

    Holmes et al. (2003) did include some variable for maternal seafood consumption in their calculations.

    But to my knowledge (and I could be wrong), none of the hair mercury studies controlled for, or reported dietary consumption ratios of protein to fat and carbs. Remember though, this would not necessarily be an apples to apples comparison with Holmes et al. being first baby haircuts, and looking at maternal diet.

12. Comment by culvercitycynic — 18 November, 2007 @ 9:39 pm

    What got me looking in that direction (protein consumption ratios) had to do with some of those outliers and other odd things … the numbers are just too all over the place. Since hair is made up of protein and it’s really not that reliable (as I understand) in terms of measuring excretions, wouldn’t those who consume more protein have more of ‘whatever’ in their hair than those who don’t consume as much protein?

13. Comment by Do'C — 18 November, 2007 @ 10:00 pm

    If you’re getting at dietary influence on hair formation (growth rates, etc.), I would think protein consumption (among other possibilities) would certainly have some impact, but, despite being bald myself, this is an area I have not read much about.

14. Comment by daedalus2u — 19 November, 2007 @ 12:27 pm

    In the body, mercury is conjugated to thiol containing amino acids and transported as the mercury-thiol complex, including a glutathione-mercury complex. These thiol containing amino acids are essential amino acids, and hair has a large thiol content. These mercury containing thiol complexes are transported pretty much “the same” as the non-mercury containing amino acid. The mercury content of hair reflects the ratio of mercury containing thiols to non-mercury containing thiols when the hair was formed. I discuss some of this in my blog on mercury physiology.

    http://daedalus2u.blogspot.com/2007/03/discussion-of-false-mercury-causes.html

    There are many studies that show a very good correlation between mercury in hair and mercury in blood. This study (n=838) finds the mercury content in hair to be 342 +/- 20 times that of blood in children aged 1-5.

    http://www.ehponline.org/members/2004/7046/7046.html

    The major excretion pathway for mercury is in the stool. It turns out that conjugation of glutathione to various chemical species is a major detoxification pathway for both normal and xenobiotic chemicals in the body including catecholamines. These glutathione conjugated species are then excreted in the bile which goes into the intestines where bacteria metabolize the mercury conjugates, forming metallic mercury which binds to sulfur or a porphyrin and is not reabsorbed.

    Mercury doesn’t have its own excretion pathways, it simply tags along with the thiol containing amino acids, and especially glutathione. Conjugation to glutathione is a major detoxification pathway. Something like 10% of the soluble proteins in rat liver are glutathione S-transferases that catalyze the attachment of glutathione to “stuff” so it can be excreted in the bile. It is gradients in glutathione concentration that provide the driving force for the transport of many compounds into the liver and into the bile where they are excreted.

    For the “low mercury excretion” idea to be correct and of a magnitude to cause mercury toxicity with hair mercury levels of less than 0.25 ppm, the ratio of hair mercury to blood mercury would need to be less than 1, rather than 342 +/- 20. An “outlier” of that magnitude (300 times) would be obvious, surprising, and any scientist observing it would want to understand it because it would be astounding (if it were ever actually observed in any human or experimental organism).

    Thiol physiology is extremely important in many hundreds of different pathways. For thiol physiology to be perturbed by a few orders of magnitude would be remarkable, and likely incompatible with life. For “mercury efflux disorder” to be true, would require the discarding of much of what is well known of mercury physiology, thiol physiology, amino acid transport, liver metabolism, and a great many other things.

    “Mercury efflux disorder” is an astounding conclusion by a few researchers working together and based on zero actual data. If it were correct, we would have to discard as completely wrong the data and conclusions of thousands of scientists working independently over decades.

    In terms of how diet would affect this, the amino acid content of hair is pretty fixed and independent of diet. If you had such a deficiency that hair wouldn’t grow at all, you would have a lot of other problems. If hair is growing at all, it is going to reflect the ratio of mercury containing to mercury free precursor amino acids.

15. Comment by C DeSoto — 19 November, 2007 @ 2:20 pm

    Hi Bloggers.
    What an enlightening week end. We hope that “DoC,” and others are remembering that an error was discovered in another authors’ work about autism (which the original authors have determined likely occurred during typing – see their correction as well as the editor in chief’s explanation of how the error occurred, which are published in the same issue of JCN). It seems of interest that the authors who found the error (me) are being attacked for having potential ulterior motives – while no such possibility is considered for the ones who actually made the error. How is this possible? I assumed when I contacted the editor (and still assume this) it was an honest mistake by the original authors. I further assumed that everyone interested should be happy it was spotted and corrected … but as I read DoC’s diatribe, I get the uncomfortable feeling this is just not so.

    Rats. For those of you who really want to know:

    Medical Hypotheses is a scientific journal that is indexed in Science Citation Index, Index Medicus/Medline, Adonis, BIOSIS, Chemical Abstracts, Elsevier BIOBASE/Current Awareness in Biological Sciences, Current Contents/Clinical Medicine, Current Contents/Life Sciences, EMBASE/Excerpta Medica, Medical Documentation Service, Reference Update, Research Alert, SciSearch, UMI (Microfilm), Russian Academy of Science. It does have a different editorial philosophy, but does publish important and widely cited papers. The journal was founded by a world renowned scientist who published over 900 peer reviewed articles in his lifetime.

    DoC suggests that the possible similarity to autism and mercury poisoning has been disproven. I do not think this can be said. Certainly, one short article offering a different view would not be considered to prove any hypothesis incorrect. Furthermore, a quick lit search will illustrate dozens of peer reviewed articles on the topic of mercury and autism. (If DoC does not have access to the original article we cite as he claims, it would be customary to withhold criticisms of it. But, for the record, the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.)

    It is suggested by DoC that because the terms “misdiagnosis” and “mercury poisoning” do not generate article hits in a PubMed search that there are no articles relating to the idea that mercury poisoning might show some similarity to autism. I too can come up with all sorts of PubMed search term combinations that yield zero hits. I wonder if is worth noting that searching for “autism diagnosis mercury” in PubMed brings up a couple of dozen hits? Also, mercury poisoning and autism brings up dozens of hits. Why is DoC writing these things? Check it out for yourself in PubMed.

    DoC has said that it is “misleading at best” to suggest that autism is on the rise. Give me a break. I am well aware that there is a controversy about how much of the increase in autism is due to more awareness and diagnostic issues. I am also well aware of the general consensus and the range of informed scientific opinion. Conservative estimates are that actual rise is three or four fold with the rest due to changes in diagnostic practices. This issue is well covered in the literature, as I suspect DoC may well be aware. It is hardly misleading to say Autism is on the rise.

    Holmes et al is ridiculed. Let me be clear, the data support Holmes’ hypothesis. And to not say so would have been self-censorship. Holmes may be wrong, the hair excretion idea may be incorrect. When analyzing the data (at the request of the journal editor) blood levels did relate to diagnosis and the hair sample levels did not. This seemed odd to me. Further investigation about why this might be led me the Holmes article. The difference in variance for the two samples stared us right in the face in black and white. The connection to Holmes was obvious. As a scientist, I am into publishing what the data show. That’s it.

    There are some misleading statements in DoC’s section on statistics. From what Doc writes, it might seem that Ip et al printed a retraction, and then we (DeSoto and Hitlan) attacked. This is not at all what happened. The retraction by the original authors came out in tandem with our reanalysis in the November issue along with an accompanying article from the editor in Chief attempting to explain how such an important error got through the peer review process. Ip et al’s retraction was I think required by the editor. As evidence for the fact that what DoC has written will mislead those who read his musings, Interverbal states that she is anxious to see how Ip et al will respond. The response is in the journal. It is a retraction and correction. Again, an error was found, politely reported and corrected.

    It is not even remotely correct to say that most of our concerns were “are mainly statistical in nature”. Our main concern is that false, mistaken data was published. We think this is crystal clear in our article. Dr. Hitlan is a statistics expert, who has read and taught from dozens of text books that deal with the topic of how to deal with extreme outliers in a data set. Neither of us are aware of any suggestions to simply ignore them. But yes, there are different schools of thought on precisely how to deal with them. None of this changes the essentials of the article… that Ip et al’s original report was in error and that a relationship does exist for mercury and autism diagnosis in the dataset.

    So much more to say… does it seem a bit pointless to say that usually one picks the type of test (one or two tailed) before the data analysis. Since the data had already been collected…? But, of course, we did realize the test should have been one-tailed before seeing the data set itself. Is DoC suggesting that we should have planned to do a one-tailed test before reading the article?

    One hears much about the conspiracy theorists that see ulterior motives in everything. Before this, I had thought this would have been more applied to the group of scientists who publish about vaccines relating to autism (a topic I have never addressed). But, as I see the response on this blogsite… where there is no mention of anyone suggesting that Ip et al might have intentionally flubbed the analysis or that their typist intentionally flubbed it or anything – but instead the scientist who stumbles across the error and reports it to the editor who then asks the scientist to reanalyze the data prints a retraction of the original results – Why am I being attacked again?

    Finally, I recall when I first learned about statistics, I too made blanket statements about statistical significance. It reflects a black and white non-nuanced view. But – if anyone knows DoC ask him this – and MAKE HIM ANSWER.
    “DoC, the connection between mercury levels and diagnosis of autism reaches a probability of p=.056 two tailed or p = .028 one tailed. If you were a betting man—what would the odds be when betting whether mercury levels are related to autism diagnosis?
    - 6 to 10 that mercury is related?
    - 8 to 10?
    - 9 out of 10?”
    The answer is that in either case the odds are way over 90% … the question that the one- versus two-tailed might matter for is if the odds that mercury is in fact related to autism diagnosis based on this data set is either better than a 94% chance that it is a real relationship, or better than a 97% chance that it is. Is Doc really quibbling that it doesn’t matter that Ip et al falsely reported that no relationship exists?

    And… finally, yes we read all of the comments with interest. We are pleased that our article is having such an impact, but hope that those interested will read the actual article rather than relying on DoC’s or any others’ analysis of what we have said (and not said). A case in point, it would not be contrary to any of our claims to find that mercury in the hair did not differ (a 1985 article by Wecker was mentioned as finding no relation and therefore of being contradictory to our findings). Note—we did not find significant hair differences either (!); we would actually see such a result as quite in keeping with the ideas of Holmes.
    We think Blogs are great. Open forums read by all sorts of folks. However, we will not be posting again on this website, please keep informed.

    See JCN website in November if you want to see the articles—all three are pertinent, which was omitted in the original posting by DoC.
    http://jcn.sagepub.com/current.dtl

16. Comment by Do'C — 19 November, 2007 @ 3:47 pm

    Dr. DeSoto,

    First and foremost, thank you for making the time to respond to our article. I do intend to respond in much more detail, but time does not permit at the moment. Until then, I am somewhat surprised that you support this statement:

    Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered.

    With the following:

    It is suggested by DoC that because the terms “misdiagnosis” and “mercury poisoning” do not generate article hits in a PubMed search that there are no articles relating to the idea that mercury poisoning might show some similarity to autism. I too can come up with all sorts of PubMed search term combinations that yield zero hits. I wonder if is worth noting that searching for “autism diagnosis mercury” in PubMed brings up a couple of dozen hits? Also, mercury poisoning and autism brings up dozens of hits.

    I think your original use of the word “sometimes” would indicate to most readers that mercury poisoning being misdiagnosed as autism is a common occurrence to some degree.

    Your absolutely right that your searches turn up more “hits” at PubMed, but how many of them are actually relevant to your original statement? “Sometimes” does not mean “once”, or even “twice”. “Mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered” is not that same as “articles relating to the idea that mercury poisoning might show some similarity to autism”.

    That’s just plain old-fashioned equivocation.

    If you would be so kind, please cite three cases in the peer-reviewed medical literature where a documented presumptive diagnosis of DSM “autism” has been given to a patient who actually suffered from mecurism instead.

17. Comment by Ms. Clark — 19 November, 2007 @ 4:18 pm

    Well, The good Dr. Desoto blasts through here and announces that she won’t be back. The tone is decidedly unprofessional, in my opinion, and some of the “evidence” she give is absolutely hysterical.

    How on earth did she manage to get a PhD if she can look at the Holmes study and say it supports their hyptohesis?
    There are kids in the the Holmes study who have 30 to 100 times the NORMAL amounts of mercury that would be expected to be in a child’s hair! And this gets no mention, not even by the venerable Dr. DeSoto. These kids who have 30 to 100 times the normal amount of mercury in their “baby hair” somehow survive this onslaught and have NO neurological sequelae at all??? Really? How about that old dose response thing? And the autistic kids have absolutely NORMAL amounts of mercury in their hair but this is reported as being LOW. Why is that? Well maybe becaue they put the autistic kids next to the “normal” kids whose hair looks like it was soaked in thimerosal or something it’s so high in mercury, and by comparison they look “low”.

    So at the last minute while anaylzing the data, Holmes (who seems to have lied about her son’s recovery from autism, by the way) calls up the hysterically extreme Boyd Haley and asks him how this could be, and on the spot Haley invents the “low excretor” phenomenon. Never mind that no one is a “low excretor” among the samples. They all have normal to extremely high amounts of mercury in their hair.

    How can DeSoto insinuate that since Medical Hypotheses is indexed that it’s the equivalent of a peer reviewed publication, when it is absolutely not.

    As far as the epidemic goes the only conservative statment that could be conceded based on the evidence is that there **may have been** a doubling of the **most strictly** defined kind of cases of autism over the past 30 years. Going form extremely rare, to very rare. The rest of the “epidemic” is in kids who never would have been counted as autistic in the past.

    The most strictly defined kind of autism doesn’t look like MR and doesn’t look like mere quirkiness. There are still very few of those kids.

    Maybe Dr. DeSoto could tell us about the reverse epidemic where we have so many fewer kids with MR and we are “losing” kids with “childhood schizophrenia” and “minimal brain dysfunction” and a dozen other diagnoses that have disappeared entirley, and how that’s paralleled this “increase” in “autism.”

    I still want to know if DeSoto fancies herself as an expert witness in the Omnibus litigation or not. Maybe not, but it’s a fair question.

18. Comment by terri — 19 November, 2007 @ 4:20 pm

    Give me a break? The leading epidemiologists of autism do not support the conclusion that “autism is on the rise.” That is a hypothesis only not received knowledge, and so far that hypothesis has not been accepted by the scientific community. For De Soto to suggest that the diagnostic and awareness issues are to blame for the increase in autism prevalence suggests she has not read the literature very well. Diagnostic and awareness issues are one part of a larger group of factors, including the broadening of the criteria in research studies, different epidemiological methods and diagnostic assessments that make comparison between today’s rates and past rates tantamount to comparing apples and oranges, newer phenotypes (Aspergers and PDD-NOS account for a large amount of the supposed increase) widespread diagnostic substitution, and the IDEA codings, among other things. The literature cited in De Soto paper is sophomoric literature at best. It is really disappointing to see such mediocre literature.

19. Comment by Ms. Clark — 19 November, 2007 @ 4:45 pm

    Good grief, I thought we were talking about a professor professor. One that’s been around the block. Maybe sort of like this professor
    http://www.dcip.org/?q=img_assist/gen/138

    No.
    http://www.uni.edu/desoto/

20. Comment by Joseph — 19 November, 2007 @ 5:27 pm

    Conservative estimates are that actual rise is three or four fold with the rest due to changes in diagnostic practices. This issue is well covered in the literature, as I suspect DoC may well be aware.

    Dr. Desoto: What are you referring to here? There is no epidemiological data that I’m aware of that could be compared and determined to be methodologically equivalent over time such that an actual rise may be inferred.

21. Comment by Interverbal — 19 November, 2007 @ 5:51 pm

    Dr. DeSoto,

    As one of the authors of the post, thank you for taking the time to comment. I quote you frequently below, in an attempt to try to address some of your points. Before I begin I want to first note that our post was not an attempt to set Ip et al. on a pedestal and knock yourself and Dr. Hitlan into the mud. We acknowledge that Ip et al. made errors. I am pleased you took the time to catch these errors. I also don’t see this as an issue of honesty in either direction. Do’C and I disagreed with several aspects of your critique and that is what our post was about.

    “What an enlightening week end. We hope that “DoC,” and others are remembering that an error was discovered in another authors’ work about autism (which the original authors have determined likely occurred during typing – see their correction as well as the editor in chief’s explanation of how the error occurred, which are published in the same issue of JCN).”

    We do remember it. We have noted the editor’s comments as well as actual data. Our critique is not that you found an error, but in other aspects of the article.

    “It seems of interest that the authors who found the error (me) are being attacked for having potential ulterior motives – while no such possibility is considered for the ones who actually made the error. How is this possible? I assumed when I contacted the editor (and still assume this) it was an honest mistake by the original authors. I further assumed that everyone interested should be happy it was spotted and corrected … but as I read DoC’s diatribe, I get the uncomfortable feeling this is just not so.”

    I do not believe have ulterior motive. I am pleased you found an error.

    “Medical Hypotheses is a scientific journal that is indexed in Science Citation Index, Index Medicus/Medline, Adonis, BIOSIS, Chemical Abstracts, Elsevier BIOBASE/Current Awareness in Biological Sciences, Current Contents/Clinical Medicine, Current Contents/Life Sciences, EMBASE/Excerpta Medica, Medical Documentation Service, Reference Update, Research Alert, SciSearch, UMI (Microfilm), Russian Academy of Science. It does have a different editorial philosophy, but does publish important and widely cited papers. The journal was founded by a world renowned scientist who published over 900 peer reviewed articles in his lifetime.”

    I am sure that all of that is true. But is still not peer reviewed. And while peer review is not foolproof, both Do’C and I are preferential towards peer review.

    “DoC suggests that the possible similarity to autism and mercury poisoning has been disproven. I do not think this can be said. Certainly, one short article offering a different view would not be considered to prove any hypothesis incorrect. Furthermore, a quick lit search will illustrate dozens of peer reviewed articles on the topic of mercury and autism. (If DoC does not have access to the original article we cite as he claims, it would be customary to withhold criticisms of it. But, for the record, the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.)”

    I think it is reasonable to criticize this study on the single fact that the diagnosis was made at such a young age (11 months).

    “DoC has said that it is “misleading at best” to suggest that autism is on the rise. Give me a break. I am well aware that there is a controversy about how much of the increase in autism is due to more awareness and diagnostic issues. I am also well aware of the general consensus and the range of informed scientific opinion. Conservative estimates are that actual rise is three or four fold with the rest due to changes in diagnostic practices. This issue is well covered in the literature, as I suspect DoC may well be aware. It is hardly misleading to say Autism is on the rise.”

    Where is such an opinion in the autism epidemiology? I may well have missed it, if so, will you please provide some references and/or quotes that would help a couple of bloggers better understand the issue? Also, will you please let me know if PDD-NOS was included in the “before” data. PDD-NOS currently constitutes roughly 75% of the prevalence. Not many people seem to understand or appreciate this.

    “There are some misleading statements in DoC’s section on statistics.”

    The statistics portion was my contribution. I welcome your criticisms and comments.

    “From what Doc writes, it might seem that Ip et al printed a retraction, and then we (DeSoto and Hitlan) attacked.”

    I was not attempting to imply this. I appreciate the chance to clarify.

    “Interverbal states that she is anxious to see how Ip et al will respond. The response is in the journal. It is a retraction and correction. Again, an error was found, politely reported and corrected.”

    Noted, however, I was thinking more about any further comment from Ip et al. on the your reanalysis, especially the statistical portion.

    “It is not even remotely correct to say that most of our concerns were “are mainly statistical in nature”. Our main concern is that false, mistaken data was published. We think this is crystal clear in our article.”

    From my reading of your article, I didn’t look like this to me. This is might be understandable since a healthy portion of your article focuses on statistical aspects. However, since you have clarified, consider this retracted.

    “Dr. Hitlan is a statistics expert, who has read and taught from dozens of text books that deal with the topic of how to deal with extreme outliers in a data set. Neither of us are aware of any suggestions to simply ignore them. But yes, there are different schools of thought on precisely how to deal with them.”

    Whereas I am not a statistics expert, not even close. I welcome any further explanation on this matter that you or Dr. Hitlan care to make. Maybe a good start, would be a discussion over how to deal with outliers?

    “None of this changes the essentials of the article… that Ip et al’s original report was in error and that a relationship does exist for mercury and autism diagnosis in the dataset.”

    You are right, but neither Do’C nor I argued against this.

    “So much more to say… does it seem a bit pointless to say that usually one picks the type of test (one or two tailed) before the data analysis. Since the data had already been collected…? But, of course, we did realize the test should have been one-tailed before seeing the data set itself. Is DoC suggesting that we should have planned to do a one-tailed test before reading the article?”

    That is a fair rebuttal. If you realized that the test should have been one-tailed before reading the data, then my point doesn’t stand.

    “Why am I being attacked again?”

    You were not attacked in our article.

    “Finally, I recall when I first learned about statistics, I too made blanket statements about statistical significance. It reflects a black and white non-nuanced view. But – if anyone knows DoC ask him this – and MAKE HIM ANSWER.”

    Since I wrote this portion you can ask me. I promise to answer.

    “DoC, the connection between mercury levels and diagnosis of autism reaches a probability of p=.056 two tailed or p = .028 one tailed. If you were a betting man—what would the odds be when betting whether mercury levels are related to autism diagnosis?
    - 6 to 10 that mercury is related?
    - 8 to 10?
    - 9 out of 10?”

    “The answer is that in either case the odds are way over 90% … the question that the one- versus two-tailed might matter for is if the odds that mercury is in fact related to autism diagnosis based on this data set is either better than a 94% chance that it is a real relationship, or better than a 97% chance that it is. Is Doc really quibbling that it doesn’t matter that Ip et al falsely reported that no relationship exists?”

    But… they didn’t falsely report that no relationship existed using a one tailed test. They used a two tailed test. As to a non-nuanced view statistics; I can definitely potentially accept that I have an inadequate view of statistics. However, help me overcome such limitations, tell me when a statistically insignificant score is close enough so that we can treat it like a statistically score. What is good enough here and how do we know?

    “We think Blogs are great. Open forums read by all sorts of folks. However, we will not be posting again on this website, please keep informed.”

    Thank you for your time.

22. Comment by Joseph — 19 November, 2007 @ 6:39 pm

    The fact that Ip et al. published errata is a point in their favor. It’s indeed a noble endeavor to go and find errors in published data. When errors are found in one’s work (be it papers or blog posts) the intellectually honest thing to do is to publish errata.

    However, I find it strange that Dr. Desoto would not appear to see a problem in the data from Holmes et al. for example, where the controls have considerably high levels of mercury in their hair relative to published reference values.

    If Dr. Desoto is inclined to scrutinize papers, why not scrutinize, say, papers by Geier & Geier? Take Geier & Geier (2004) ( http://www.medscimonit.com/fulltxt.php?ICID=11608 ) for example. See figure 2, particularly the years in the X axis (hint: it skips from 1985 to 1990). This is the kind of thing that cannot be explained as a simple statistical error, and I have never heard of Geier & Geier publishing errata.

    In the case of Ip et al. I don’t believe there’s enough to say the authors intentionally altered data, as has been insinuated. Furthermore, I don’t see that the errors found alter the conclusions of the paper in any significant way.

23. Comment by Do'C — 19 November, 2007 @ 8:15 pm

    What an enlightening week end. We hope that “DoC,” and others are remembering that an error was discovered in another authors’ work about autism (which the original authors have determined likely occurred during typing – see their correction as well as the editor in chief’s explanation of how the error occurred, which are published in the same issue of JCN).

    I’ve seen the errata and the editor’s comments. Although never in dispute, it is noted that an error was discovered in the other authors’ original work about autism.

    It seems of interest that the authors who found the error (me) are being attacked for having potential ulterior motives – while no such possibility is considered for the ones who actually made the error.

    Motives (ulterior or otherwise) were neither mentioned, nor discussed in our article. If you are referring to comments posted by others, you may certainly address them directly.

    How is this possible? I assumed when I contacted the editor (and still assume this) it was an honest mistake by the original authors. I further assumed that everyone interested should be happy it was spotted and corrected … but as I read DoC’s diatribe, I get the uncomfortable feeling this is just not so.

    I’m happy you spotted the mistake and pursued correction. Please don’t feel uncomfortable. Look at this as a potential opportunity to re-examine some of the science you’ve chosen to rely on for your paper, as well as your knowledge about anatomy and physiology. I’ll re-examine mine as well.

    Rats. For those of you who really want to know:

    Medical Hypotheses is a scientific journal that is indexed in Science Citation Index, Index Medicus/Medline, Adonis, BIOSIS, Chemical Abstracts, Elsevier BIOBASE/Current Awareness in Biological Sciences, Current Contents/Clinical Medicine, Current Contents/Life Sciences, EMBASE/Excerpta Medica, Medical Documentation Service, Reference Update, Research Alert, SciSearch, UMI (Microfilm), Russian Academy of Science. It does have a different editorial philosophy, but does publish important and widely cited papers. The journal was founded by a world renowned scientist who published over 900 peer reviewed articles in his lifetime.

    But it is still a pay-to-publish, not peer-reviewed journal. Like Interverbal, while imperfect for sure, I am preferential towards peer-review.

    DoC suggests that the possible similarity to autism and mercury poisoning has been disproven. I do not think this can be said. Certainly, one short article offering a different view would not be considered to prove any hypothesis incorrect.

    A hypothesis should be proven correct. If you want to argue that autism and mercury poisoning are similar, the burden of proof (not the provision of hypothesis) is yours. To be clear, I am not saying that the hypothesis has been disproven, I am saying that Bernard et al. (2001) was invalidated by Nelson and Bauman (2003), ergo you’ll still need to bring some proof.

    Furthermore, a quick lit search will illustrate dozens of peer reviewed articles on the topic of mercury and autism.

    If you think any in particular are relevant to your paper, please feel free to present them.

    (If DoC does not have access to the original article we cite as he claims, it would be customary to withhold criticisms of it. But, for the record, the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.)

    I do not have access to the original article you cited. I also think it is reasonable to criticize this study on the single fact that the diagnosis was made at such a young age (11 months).

    It is suggested by DoC that because the terms “misdiagnosis” and “mercury poisoning” do not generate article hits in a PubMed search that there are no articles relating to the idea that mercury poisoning might show some similarity to autism. I too can come up with all sorts of PubMed search term combinations that yield zero hits. I wonder if is worth noting that searching for “autism diagnosis mercury” in PubMed brings up a couple of dozen hits? Also, mercury poisoning and autism brings up dozens of hits. Why is DoC writing these things? Check it out for yourself in PubMed.

    Addressed in a previous comment.

    DoC has said that it is “misleading at best” to suggest that autism is on the rise. Give me a break. I am well aware that there is a controversy about how much of the increase in autism is due to more awareness and diagnostic issues. I am also well aware of the general consensus and the range of informed scientific opinion. Conservative estimates are that actual rise is three or four fold with the rest due to changes in diagnostic practices.

    Evidence please!!!

    This issue is well covered in the literature, as I suspect DoC may well be aware. It is hardly misleading to say Autism is on the rise.

    Then you will have no trouble demonstrating this increase in actual autism prevalence with appropriate scientific support. I will look forward to it.

    Holmes et al is ridiculed. Let me be clear, the data support Holmes’ hypothesis.

    I think you may be forgetting that big “IF” at the beginning of their hypothesis. Please feel free to address it with appropriate scientific support.

    And to not say so would have been self-censorship. Holmes may be wrong, the hair excretion idea may be incorrect. When analyzing the data (at the request of the journal editor) blood levels did relate to diagnosis and the hair sample levels did not. This seemed odd to me. Further investigation about why this might be led me the Holmes article. The difference in variance for the two samples stared us right in the face in black and white. The connection to Holmes was obvious. As a scientist, I am into publishing what the data show. That’s it.

    In my humble opinion, I think you may have accepted the “poor excretor” hypothesis uncritically. It could very well turn out to be true, but the data do not support it without an explanation of that “IF”. The data show lower mercury in the circulating blood at the time the hair follicles were formed, that’s it.

    …

    One hears much about the conspiracy theorists that see ulterior motives in everything. Before this, I had thought this would have been more applied to the group of scientists who publish about vaccines relating to autism (a topic I have never addressed). But, as I see the response on this blogsite… where there is no mention of anyone suggesting that Ip et al might have intentionally flubbed the analysis or that their typist intentionally flubbed it or anything – but instead the scientist who stumbles across the error and reports it to the editor who then asks the scientist to reanalyze the data prints a retraction of the original results – Why am I being attacked again?

    You were not attacked. Some of the potential credulity of statements in the article, and the science relied upon, was criticized.

    However, we will not be posting again on this website, please keep informed.

    That’s really a shame because along with your pointing out of an error, your participation in the discussion here is appreciated. If you do venture back, I’d personally like to hear why you seem to find Holmes et al. so compelling, especially if you have a chance to re-read it along with several of the criticisms of it.

24. Comment by Do'C — 19 November, 2007 @ 11:15 pm

    I pulled a copy of Chrysochoou et al. tonight.

    Interesting actually. Definitely not autism. Remember what DeSoto wrote:

    Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered. [4]

    That’s not exactly what the paper says.

    “The diagnosis remained obscure. The child was referred for further evaluation of severe psychomotor regression with autistic features of unknown aetiology.”

    That’s not a “diagnosis” of “autism” in my opinion. In reality this child presented with so many physical symptoms that don’t look anything like autism, it’s surprising to me that Dr. DeSoto would imply that diagnosis was “autism of unknown etiology”.

    But, for the record, the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.)

    The paper did not mention a “preliminary diagnosis”. It mentioned a referral “for further evaluation of severe psychomotor regression with autistic features of unknown aetiology”.

    Would you like to retract your claim, from the record Dr. DeSoto?

    “Arterial hypertension responded well to b-blockers, neurological deficits improved, the skin rash cleared slowly, and he could be discharged 2 weeks later. Regular follow-up confirmed complete remission of symptoms and catchup growth; at the age of 2 years his development is normal and he remains in good health.”

    Incidentally, the child is no longer has “autistic features” to go with the psychomotor regression apparently. What a load of horseshit to imply that “autism of unknown etiology” was later shown to be due to mercury exposure (as if it were still autism, but due to mercury exposure).

    I can’t wait to see those three citations.

25. Comment by Joseph — 20 November, 2007 @ 7:30 am

    Looks like Dr. Desoto did not read her citation #4. Either that, or she misrepresented it. I hope we’ll see a retraction of the relevant statement from her.

    There are a lot of real shenanigans going on in autism research, all the way to plagiarism. And it’s not the scientists who do publish errata who are the problem in my view.

    If what DoC just posted is in dispute, I guess we could contact Chrysochoou et al. and ask them if “autism of unknown etiology” was a proper characterization of their reporting.

26. Comment by daedalus2u — 20 November, 2007 @ 7:48 am

    I think Ms Clark has it pegged, getting “credentials” to be able to testify for $250/hr. But what with the scarcity of “experts” on mercury and autism, maybe the law of supply and demand is driving up those prices. Wouldn’t want to run out of “credible” experts before the Autism Omnibus proceedings have run their course. I suspect that is why she won’t be back to post again, wouldn’t want to give the opposition anything with which to impeach her testimony now would we? It is interesting that Dr DeSoto’s main line of research is in the effects of steroids on neurodevelopment, including testosterone and estrogen. I wonder if she is getting funding from the manufacturer of Lupron? Of course that would be totally and completely and utterly unrelated to autism and mercury, so it wouldn’t need to be disclosed.

    A single paper can disprove a hypothesis. A single fact can disprove one also.

27. Comment by mayfly — 20 November, 2007 @ 7:55 am

    I must agree attacks on character are out of place.

    I do have access to the European Journal of Pediatrics, and in their paper, “An 11-month-old boy with psychomotor regression and auto-aggressive behaviour”, Chrysochoou and her co-authors state: “The child was referred for further evaluation of severe psychomotor regression with autistic features of unknown aetiology.”which is certainly different than “autism of unknown aetiology”.

    The autistic features: included stereotypical hand movements, a sad apathetic countenance, indifference to his surroundings, and auto-aggressive behavior (biting his hands and feet). [He also bit other objects]

    The child also was suffering: developmental regression, he could no longer crawl or refuse to so, an elevated heart rate (160-180 bpm), had become rather underweight, had swollen extremities, “skin desquamation”, “axial hypotonia”, and what the authors term “brisk reflexes”. The child also had a skin rash.

    Like many such articles, different diagnoses are proposed and dismissed, and the eventual finding of mercury poisoning. The source being a broken thermometer.

    The child was treated and in two weeks was released. “Regular follow-up confirmed complete remission of symptoms and catch-up growth; at the age of 2 years his development is normal and he remains in good health.”

    In their summing up the authors state “…several disorders with a grim prognosis were ruled out before considering mercury intoxication,”

    One part of the article did give me pause was a brief discussion of mercury-containing calomel and that only about 1 in 500 children developed acrodynia from it. The authors state “It is possible that a genetic predisposition plays a role in determining whether symptoms develop after exposure to small quantities of mercury.”

    I was totally unfamiliar with calomel which is mercurous chloride Hg2Cl2 The drug goes a long way back. Mercurous chloride can change to mercuric chloride HgCl2, a much more toxic compound on exposure to sunlight.

    The drug has a very long history. I don’t know if it was given to children.

28. Comment by daedalus2u — 20 November, 2007 @ 9:24 am

    What mercury level was measured in the child to diagnose mrecury poisoning?

    Mercury was commonly used as a therapeutic treatment. Check out page 262 of this.

    Link

    On page 269, it is recommended to administer calomel (1/10 grain = 6500 micrograms) every hour until acute symptoms of mercury poisoning are reached.

    You can download the pdf

    Here

29. Comment by Prometheus — 20 November, 2007 @ 11:36 am

    In the first section of the Chrysochoou et al paper, they describe the clinical findings of the child in question at presentation:

    “Clinical examination revealed swollen hands and feet with skin desquamation, axial hypotonia and brisk reflexes. The child sweated profusely, refused to crawl or stand, showed stereotypic movements of the hands (kneading) and repeatedly bit objects or his own hands.”

    I suppose that someone not familiar with autism could be forgiven for reading “stereotypic movements of the hands” and not notice the qualifier - “(kneading)” - which showed it to be rather different from what is “typically” seen in autism (i.e. “flapping”).

    A person unfamiliar with autism (and medicine in general) might also think that a desquamating (peeling) rash and swollen hands and feet are just “red herrings” when, in fact, they are a valuable clue that this was not autism.

    In fact, someone who had not read the paper - but perhaps had just read the abstract - might think that the diagnosis of autism was made in this case, rather than simply a single comment that the child showed “autistic features of an unknown aetiology”.

    I think that it’s time for Dr. DeSoto to issue a “clarification”.

    Prometheus

30. Comment by Do'C — 20 November, 2007 @ 11:38 am

    Hmmm. There have been several visits to the blog from the University of Northern Iowa this morning. Nary a comment about any of this from Dr. DeSoto though.

    Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered. [4]

    Again, that’s not exactly what the paper says.

    “The diagnosis remained obscure. The child was referred for further evaluation of severe psychomotor regression with autistic features of unknown aetiology.”
    

    That’s not a “diagnosis” of “autism” in my opinion. In reality this child presented with so many physical symptoms that don’t look anything like autism, it’s surprising to me that Dr. DeSoto would imply that diagnosis was “autism of unknown etiology”.

    But, for the record, the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.

    The paper did not mention a “preliminary diagnosis”. It mentioned a referral “for further evaluation of severe psychomotor regression with autistic features of unknown aetiology”. In fact, that is the single mention of “autism” or “autistic” in the entire paper.

    If you’re reading Dr. DeSoto, I’ll ask again, would you like to set “the record” straight? I think doing so both here, and at the Journal of Childhood Neurology would be the ethical thing to do.

    What a load of horseshit to imply that “autism of unknown etiology” was later shown to be due to mercury exposure (as if it were autism, but due to mercury exposure).

    I’d really like to see those three citations to support your statement that mercury poisoning has “sometimes” been presumptively “diagnosed” as autism of unknown etiology until the mercury poisoning has been uncovered.

    I think you owe it to the subscribers of the Journal of Childhood Neurology to provide citations that support your assertion - or a corrected statement, such as:

    “Furthermore, although an autism diagnosis was not made, in one case report, mercury poisoning was referred for further evaluation as “severe psychomotor regression with autistic features of unknown etiology”, until the mercury poisoning was uncovered. [4]“

31. Comment by Patrick — 20 November, 2007 @ 12:29 pm

    Pardon me for piling on here but as an amateur literature analyst…

    “Furthermore, mercury poisoning has sometimes…” I take exception to this comment as their reference is to a single case study, not a review of multiple cases. Truth in advertising gone astray again! (Since their paid presentation in the Hypothesis publication is in my opinion merely advertising their questionable quality research.)

    “Because there has been a several-fold increase in environmental mercury exposure…” Unsupported statement, while I have no doubt it may be true, they need to reference the publication(s) that support this claim.

32. Comment by Do'C — 20 November, 2007 @ 1:28 pm

    (Since their paid presentation in the Hypothesis publication is in my opinion merely advertising their questionable quality research.)

    Note: It was Bernard et al. that appeared in Med. Hypotheses.

33. Comment by Ms. Clark — 20 November, 2007 @ 2:18 pm

    Dr. DeSoto needs to apologize to the editors of J. Child. Neurology, in my opinion.

    She either wasted their time by making them read a bunch of bunk disguised as intellectually driven concern or she’s just playing games here. I would like to know how much contact Dr. DeSoto had with Dr. Mark Geier prior to deciding to look closely at the Ip et al., paper. Dr. DeSoto has no expertise in autism at all, nor toxicology or epidemiology, obviously. She’s a researcher in Borderline PD (hey, Dr. DeSoto, check out the mercury moms and dads, though they might be more NPD, some of the express frightening amounts of rage if you reject them) and the effect of estrogen on it. She’s a researcher in the effects of testosterone (ding! ding! ding!) and it’s effect on personality or motivation or something. But she just happens to pick up this old paper by Ip and starts doing the math, and picks up this old paper by Holmes and totally fails to see a gargantuan error in it. Amazing.

    I think DeSoto’s supervisors need to be notified of her extremely sloppy work so they can look at her other papers to see if they are similarly bad.

    I think the Chrysochoou paper (she didn’t expect anyone to read it, right?) might be titled, Totally Reversable Mercury Vapor Poisoning as a Novel form of temporary Lesch-Nyhan Syndrome, since Lesch-Nyhan is specifically named as on of the disorders they tried to rule out.

    The thing wtih calomel and teething powder is that no one was measuring the dosages that the babies were getting. So the ones that developed Pinks Disease were logically the ones that got higher doses per body weight over longer periods of time. There is no known disease of an inability to excrete mercury. There is one for copper. Again, no one has ever found a disease where a person is unable to excrete mercury.

    But look how the baby with the high exposure to mercury didn’t become autistic at all as a result. He was sickly and weak and his hands hurt (hello! acryodynia!) but he wasn’t autistic. He bounced back really quickly (without chelation or Lupron injections for the rest of his life, either!).

34. Comment by Prometheus — 20 November, 2007 @ 2:37 pm

    Ms. Clark,

    Since it was the Editor in Chief (Roger Brumback) of The Journal of Child Neurology who contacted Dr. DeSoto to re-examine the Ip et al data, I would not expect her to apologize to him for her errors.

    The “corrected” Ip et al data showed a p = 0.06 for comparison between the autistic and non-autistic blood mercury, which is not statistically significant, but is “close”. Not doubt, choosing a different statistical analysis could potentially push that “over the edge” to signficance.

    What was not mentioned is that the data for hair mercury (ala Holmes et al) shows a p = 0.79 (really not significant).

    So, I fail to see how this validates Holmes et al or, for that matter, how the Holmes et al data “supports their hypothesis”. As I’ve pointed out before (so many times), their hypothesis isn’t supported by dozens of studies that have shown that the hair mercury content follows blood mercury content at the time the hair is formed.

    If autistic children are “poor excretors” (no data supporting that hypothesis, either), their blood mercury should be higher, right? Then their hair mercury should be higher.

    Unless and until somebody shows that autistic children have some sort of alteration in their hair physiology that alters what has been shown in several mammal species, the “poor excretor” remains unsupported by data.

    Prometheus

35. Comment by isles — 20 November, 2007 @ 3:37 pm

    Well? Dr. DeSoto? Despite your insistence to the contrary, it seems likely that you are reading.

    Can you or can’t you present evidence for the proposition that there is an active process of mercury excretion into hair which could be impaired in certain individuals?

    One more question. To what extent, if any, was your paper influenced or suggested by Mark Geier, whose work is profiled in the “SIgnificant Misrepresentations” series at neurodiversity.com? See http://neurodiversity.com/weblog/article/109/lupron-geier-index

36. Comment by daedalus2u — 20 November, 2007 @ 5:17 pm

    If there is that much discordance between the hair and blood mercury levels, then either one data set or both are suspect and not reliable.

    In any case, at best all this shows is an association, not a bit of causation.

    The major source of mercury exposure is diet, from eating fish. A 3 ounce portion of canned tuna has about 35 micrograms of mercury.

    Small flucutations in mercury level, whether they are “significant” or not from a statistical viewpoint may have no clinical importance at all.

    An observation that has very high statistical signficance, is the observation that ASD individuals have larger heads. An obvious “treatment” would be to apply pressure to shrink their heads back to “normal” size. A recent Dilbert cartoon gave me the idea.

    http://dilbert.com/comics/dilbert/archive/images/dilbert2007111111117.gif

    Maybe that will make them “normal”? A treatment that makes as much sense as chelation in the absence of measured toxic levels of heavy metals.

37. Comment by Schwartz — 20 November, 2007 @ 10:43 pm

    Ms. Clark,

    I was laughing at first as you accuse someone of sounding unprofessional in one post, and then fire back with a complete ad hom attack in the next.

    You appear to be one of the few people here alleging things bordering on fraud. IMO, there is little evidence to support that.

38. Comment by Schwartz — 20 November, 2007 @ 11:13 pm

    Isles,

    Why you and Joseph continually bring up the Geier Red Herring argument in numerous debates completely baffles me.

    Joseph,

    Your excellent blogs and arguments around prevalence data have convinced me to remain neutral on this particular point. I would also like to see what references qualify as generally accepted increases in prevalence.

    I think there is a surprisingly (at least to much of the population) high prevalence since we’ve started to get data on this. Is it steadily increasing? I am not yet convinced of any particular answer to that question yet.

39. Comment by Do'C — 20 November, 2007 @ 11:20 pm

    Schwartz,

    Would you suggest that “misunderstandings and neglect create more confusion in this world than trickery and malice”?

    How would you explain DeSoto’s citation and representation of Chrysochoou et al.?

    How would you explain DeSoto’s assertion of “general consensus and the range of informed scientific opinion. Conservative estimates are that actual rise is three or four fold with the rest due to changes in diagnostic practices.”?

    How would you explain a statement of “The difference in variance for the two samples stared us right in the face in black and white. The connection to Holmes was obvious. As a scientist, I am into publishing what the data show.”?

    (The data show low blood mercury levels at the time of hair follicle formation, that’s it).

    You are right about ad hominem - in my opinion, it is usually counter productive. What’s your take on DeSoto’s understanding of the relevant science?

40. Comment by Schwartz — 21 November, 2007 @ 12:18 am

    Prometheus

    So, I fail to see how this validates Holmes et al or, for that matter, how the Holmes et al data “supports their hypothesis”. As I’ve pointed out before (so many times), their hypothesis isn’t supported by dozens of studies that have shown that the hair mercury content follows blood mercury content at the time the hair is formed.

    The way I’m reading the Authors comments here, is that they claim the Ip et al data supports Holmes et al hypothesis. They don’t discuss any data from Holmes et al, only the hypothesis. So a discussion of the quality of Holmes et al’s study doesn’t appear to be relevent to the specific claims by DeSoto et al.

    All of your other arguments about physiological theories and even Holmes et al study quality don’t change their argument. Given the extreme limits of the stated scope of their study, the only real debate is around the statistics and conclusions of the Ip et al study.

41. Comment by Schwartz — 21 November, 2007 @ 12:24 am

    Interverbal, Do’C

    I was misled by the wording you used to describe the errata acknowledgment by the original authors. It did sound like this paper was frivilous in pointing out an error that had already been dealt with.

    The fact that they were both published in the same journal with an explanation by the editor at the same time makes a difference, and you didn’t point that out.

    It doesn’t make any factual difference, but it certainly changes the tone of the discussion.

    More of my comments on this high quality dialog (well, mostly high quality):

    Do’C

    I do not have access to the original article you cited. I also think it is reasonable to criticize this study on the single fact that the diagnosis was made at such a young age (11 months).

    I will also support Dr DeSoto in her assertion that it would be customary to read a full reference before commenting on it. This is regardless of the eventual outcome of the debate on that reference.

    You were not attacked. Some of the potential credulity of statements in the article, and the science relied upon, was criticized.

    Dr DeSoto was certainly attacked on this blog.

    But, as I see the response on this blogsite…

    She is referring to the comments, not your post. And Ms. Clarks’ (and even Daedulus2u) clear allegations of conspiracy (bordering on fraud) certainly qualify as an attack.

    Interverbal,

    I am sure that all of that is true. But is still not peer reviewed. And while peer review is not foolproof, both Do’C and I are preferential towards peer review.

    While you may feel this way, the joint article published with your name effectively mocked the publication. It did not state a mere preference for peer-review.

    Wow! DeSoto and Hitlan seem to have fallen for that one hook, line, and stinker. Why not cite more recent, and peer-reviewed literature? (Med. Hypotheses is a pay-to-publish, not peer-reviewed journal).

    This could also be construed as an attack of character so her comments on that topic appear justified.

    Do’C

    The diagnosis remained obscure. The child was referred for further evaluation of severe psychomotor regression with autistic features of unknown aetiology.

    I could easily read a referral with a classification of symptoms of autism to mean a preliminary diagnosis so I don’t think this particular comment is as egregious as you espouse.

    However, I feel you are fully justified and correct in pointing out that this reference does not effectively support the statement in the study:

    Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered. [4]

    This quote from the study is definitely misleading given the strength of the reference to back it up.

    Do’C

    Her comments about PubMed searches are in direct response to your using a specific lack of hits to support one of your arguments. I did the queries suggested in her response, and a couple of the hits appear to be potentially relevent, but I don’t get free access to studies, so I can’t validate any of the claims. I don’t trust abstracts.

    Interverbal, Do’C
    If I were to summarize my current position:

    1) I think the claims of growing incidence of Autism in the rebuttal needs some sort of reference. My position as I stated earlier to Joseph, is completely undetermined at this point given the data I have seen to date.

    2) I think the claims that [symptoms of mercury poisoning and Autism are similar to the degree they are often mis-diagnosed] to be weakly supported by evidence and I would need to see more conclusive evidence before being convinced of this.

    3) I think the original article had an undertone of mocking, and the accusations of attack were slightly justified — they were certainly justified based on some of the commentors — but that should not detract from what I see as a relatively objective blog entry. I commend Do’C for minimizing the mocking and Interverbal for the his eternal politeness.

    4) I think that any discussion of Holme et al data or study is completely irrelevant as the argument presented here by Dr DeSoto is that Ip et al data supports Holme’s hypothesis.
    The authors appear to scrictly limit the scope of their discussion to this. And here we come back to the black art of statistics.

    It appears clear that Ip et al made a calculation or transcription error. It also appears clear that the hypothesis did not appropriately match the statistical analysis performed.

    Now we come to the final (and most important) argument: that the results don’t match the conclusion. And the verdict here lies in subjective decisions about interpreting the data.

    We have Interverbal who is clearly black and white when it comes to analysis and we have the authors who argue that the data from the Ip et all study shows support for Holme’s hypothesis contrary to the conclusion of the authors.

    The points of contention being treatment of outliers, but mainly interpretation of significance in the results and the resulting conclusion.

    I am far less an expert in statistics, so I can’t really draw a strong conclusion here. But I guess I’ll be dishing out another $25.00 to get access to this study.

42. Comment by Schwartz — 21 November, 2007 @ 12:45 am

    Do’C

    Would you suggest that “misunderstandings and neglect create more confusion in this world than trickery and malice”?

    

    I do not get impression that this study is based on trickery and malice. Are there problems? Yes, I think there are.

    I was writing my long response while you posted that last one to me. I think I answered most of your questions in my other post. Although I agree with a few points in her rebuttal, I substantially agree with your position on both the lack of evidence supporting the general statements on increase in Autism prevalance, and the question of Autism symptoms being similar to Mercury poisoning also appears to lack the appropriate references.

    I may not agree with the exact path that you got there in the latter instance, but I fully support the substance of your argument.

    I think where the overall argument runs into a stalemate is that the authors have limited their scope of discussion to the results of Ip et al data supporting Holmes et al Hypothesis. And that all comes down to interpretation of the statistics.

    DeSoto is pretty clear that they aren’t arguing that the Holmes et al Hypothesis is correct, only that the Ip et al data supports it.

    What’s your take on DeSoto’s understanding of the relevant science?

    I think I have to read the study now before I pass final judgement on that one.

    I think you and Interverbal have presented good arguments and I’m grateful for the effort you’ve put into this.

    You are right about ad hominem - in my opinion, it is usually counter productive.

    I personally find the Geier Red Herring arguments increasingly tiring and distracting as well.

43. Comment by HCN — 21 November, 2007 @ 12:51 am

    Meandring the blogosphere (it started with Scienceblogs where I went to PZ Myers page on to a posting he made about the Discovery Institute plagiarizing a video on cell processes by Harvard noticed by a graduate student name Abbie Smith studying HIV, and her debate with a denialist dentist), I came upon an interesting assessment of the journal called “Medical Hypothesis”:
    Translation: “We’re a journal for conspiracy theorists to publish in to proclaim they have a published peer reviewed article.”

    From http://endogenousretrovirus.blogspot.com/2007/03/why-horowitz-published-in-medical.html

44. Comment by Ms. Clark — 21 November, 2007 @ 2:16 am

    What I find amusing is that if I try to stand toe to toe with a PhD, which I have attempted, actually, practically nose to nose, and hold forth an opinion that is likewise held by a few dozen other easily nameable PhDs and MDs, I will be told that I have no standing in the discussion as I am not a professional.

    No, I am a plebian. I have been (virtually) spat upon by certain PhDs, well, a couple of them anyway, who didn’t like what I was saying about their work (again it was not just my opinion, it was the opinion of a cadre of PhD’s and MDs I had discussed it with).

    So if I can’t be a “professional” in rank (that’s fair), I don’t see why I need to behave in a “professional” manner. Dr. Mary Catherine shouldn’t be coming here and throwing a snit fit. (my opinion) Her paper we are discussion is juvenile in tone, in my opinion, and even (!) includes statements (!) with exclamation (!) points in them. Which (!) I don’t remember ever seeing in a peer reviewed published work before! I certainly would not have put 2 or even one exclamation point in any of the science papers I wrote as a student while at UCD. I think it would have been seen as ridiculous and I would have been told that professionals don’t use them in peer reviewed papers… but I digress.

    To be clear, I asked a pertinent question. If you see how the expert witnesses for the petitioners in the Omnibus hearing have been used, courted, paid, etc, and how they sometimes begin by jumping in the fray from NOWHERE with a paper, a very weak paper, but a paper nonetheless on something related to the topic at hand. I think it would be stupid not to ask Dr. DeSoto if she intended to be a witness in the Omnibus hearings. I find it fascinating that so many feel no one has a right to ask. I did not say, “for a fact the woman is on the take.” I personally have been told “for a fact” that **I** am in someone’s pay, maybe 20 or 30 times over the past few years. All of it was off base and unfounded. At any rate, maybe since I am so frequently the object of such questioning that it seems more or less normal to me.

    But get real people. This is how the vaccine court game is played. They find marginal “experts” and pay them a lot of money to testify to whatever is needed to be said.

    If Dr. DeSoto is innocent, as she may well be, then all she has to do is say, “No.” It’s not like her reputation was destroyed by my asking the question here.

    And given that her previous work has been in testosterone, a sudden entry into autism surely raises the spectre of the vampire, Geier. If you don’t like that, well, that’s tough. It’s a coincindence that’s hard not to comment on. (Isles did, that one wasn’t mine.) The Geiers are the ones who apparently gave Lenny Dr. DeSoto’s paper (while it was still embargoed,I heard) to post on EoHam, by the way.

    I apologize for saying that it was Dr. DeSoto’s idea to review the data. I got that wrong. Still why would anyone ask her to review this data? Anyone got an idea? I wish someone would ask someone to review the Holmes paper officially. It’s hysterical.

    Finally, DoC doesn’t allow people to call people names on his blog, which is right. He allows people to ask questions that most people might not. I’m not calling Dr. Mary Catherine any bad names, on the other hand she came here and said “bloggers” like it was a bad thing, like we were too far beneath her. That’s unprofessional. :-]

45. Comment by Joseph — 21 November, 2007 @ 7:06 am

    Is Dr. DeSoto going to address the alleged error in her paper, yes or no? It would be extremely dishonest and ironic if she refuses to address an error in a paper of hers that was supposed to expose the errors of another paper.

46. Comment by daedalus2u — 21 November, 2007 @ 7:46 am

    Schwartz, I don’t know how familiar you are with the autism-mercury controversy and the “bad blood” there is about it. But you should look at Orac’s blog.

    Link 1

    Link 2

    Link 3

    Link 4

    Generation Rescue still touts Wakefield as one of their “heroes”. What he published about PCR and measles virus was known to be false by him when he published it. That paper was retracted by 8 of 10 authors years ago, but Generation Rescue makes no mention of that on their website. To me, that is fraud.

    Link 5

    Link 6

    Link 7

    These things are not close calls. Threatening people with violence is the work of terrorists and extortionists not well meaning advocates. Lying about research results is fraud. I think it is legitimate to judge people by those they hold in high regard. The anti-mercury crowd holds people that I know to be liars, cheats, bullies, quacks and frauds in high regard. I know them to commit fraud themselves. When they have been caught in numerous lies, it is not unreasonable to severely question everything they say.

    Is Dr DeSoto one of “them”? I don’t know. I haven’t read the paper yet, just the abstract and reviews. Citing the “mercury causes autism” idea and the “mercury efflux disorder” idea uncritically is either extremely sloppy and poor scholarship, or it is deliberate spin to boost “credibility” in legal proceedings. Wakefield was paid 700,000 pounds to generate his anti-MMR stuff by the lawyers who filed suit. Who funded Dr DeSoto’s research and evaluation of the Ip et al paper?

47. Comment by passionlessDrone — 21 November, 2007 @ 7:58 am

    Hello friends -

    I’m just a simple drone and all of this semantic jockeying has me confused. Can someone clear some things up for me?

    My statistics are poor, but are the following two statements valid interpretations of the Ip data?

    1) If we analyze mercury blood levels and ask ourselves if any change in mercury is associated with autism, the chance that a relationship exists is 94%.

    2) If we analyze mercury blood levels and ask ourselves if higher leverls or mercury are associated with autism, the chance that a relationship exists is 97%.

    Do I have it wrong here? My apologies for my generally poor statistical knowledge.

    If my understanding is correct (?) ,there is a four hundred pound gorilla in this room that very few people want to talk about. There is a lot of wordsmithing, and some dime store accusations, but this seemingly very important nugget of information, for the most part, goes undiscussed.

    Take care!

    - pD

48. Comment by Patrick — 21 November, 2007 @ 9:48 am

    Ok, my mistake for getting carried away, sorry.

    Please edit or entirely strike my previous comment. (This one can be deleted too, I will have to be more careful in my reading.)

    The section to be removed to more properly reflect the complaint is:

    Truth in advertising gone astray again! (Since their paid presentation in the Hypothesis publication is in my opinion merely advertising their questionable quality research.)

49. Comment by Do'C — 21 November, 2007 @ 10:20 am

    Joseph, I’m terribly sorry your comment was dropped. Unfortunately, a less than sign is taken as the beginning of HTML markup, and the rest of the comment truncated. I don’t have your comment as typed-in either. I hope you’ll repost (please use the words “less than”).

50. Comment by Joseph — 21 November, 2007 @ 11:42 am

    That’s alright.

    I was basically agreeing with pD that a p less than 0.06, while not considered statistically significant, does indicate that the odds a difference was not found is only 6%.

    But an association is not necessarily causal. A lot more work would be required for us to begin to suspect causality, especially in a situation like this where you would need to control for various obvious confounds, such as diet and Pica. For hair mercury, I would guess even differences in shower habits could impact the results.

    Furthermore, there is other research that points in a different direction. Didn’t even Jim Adams fail to find a difference in hair/blood mercury levels? I don’t have the reference with me.

51. Comment by Joseph — 21 November, 2007 @ 12:56 pm

    And to put in perspective the sort of minor exposure differences found in comparisons between autistics vs. non-autistics, and the significance of diet, see “Fish consumption, mercury exposure and serum antinuclear antibody in Amazonians” (Alves et al., 2006). They find that people living along the Amazon river had a mean hair-Hg level of 34ppm, whereas people living in Manaus had a mean hair-Hg of 1.0ppm.

    Imagine the scandal that would result if autistics were found to have 34 times the mercury levels in blood or hair that non-autistics have. Instead we’re left with some minor differences and discussions about statistical significance. It’s exceedingly difficult to buy causality at these low exposure levels, when the simple explanation for group differences is that the groups were different.

52. Comment by Do'C — 21 November, 2007 @ 1:09 pm

    pD,

    The P value is not the probability of the null hypothesis being true. It is the probability that the difference observed could have occurred by chance. Therefore, 1- the P value is not the probability of an alternative hypothesis being true.

    A p value greater than .05 simply says we can’t reject the null hypothesis. That’s it.

53. Comment by daedalus2u — 21 November, 2007 @ 1:40 pm

    I wonder how many negative studies of mercury in ASD individuals have not been published because they are not “interesting” to the editor?

54. Comment by Do'C — 21 November, 2007 @ 2:19 pm

    I wonder how many negative studies of mercury in ASD individuals have not been published because they are not “interesting” to the editor?

    That, is a great question.

55. Comment by Prometheus — 21 November, 2007 @ 4:32 pm

    I just wanted to amplify what Do’C had to say about statistical significance.

    The biomedical “standard” of using a p value less than or equal to 0.05 as a criteria for rejecting the null hypothesis is simply a convention. That having been said, it is a convention that has worked very well for decades.

    Rejecting the null hypothesis implies nothing more than that: the hypothesis that there is no difference between the groups can be rejected.

    However, at a p = 0.05, there is still a 5% chance that the differences seen are an artifact of sampling and that there is really no difference.

    5% is a better chance than any lottery.

    Another point that pD seems confused on is the issue of association. Finding that there is a difference between the two groups does not mean that there is a relationship between mercury and autism.

    What it does mean is that there is a statistical association between the level of one (mercury) and the presence of another (autism).

    For that matter, there is an association between shoe size and reading ability in every grade school in the country. This does not mean that you can improve your child’s reading ability by getting them bigger shoes, however.

    Prometheus

56. Comment by daedalus2u — 21 November, 2007 @ 8:01 pm

    Prometheus, you mean that clowns with the big floppy shoes are not the best doctors and the world’s greatest experts on everything?

    Someone should tell the anti-mercury crowd that. I think they are pretty hung up with correlation = causation.

57. Comment by Do'C — 21 November, 2007 @ 9:04 pm

    Schwartz wrote:

    I do not get impression that this study is based on trickery and malice. Are there problems? Yes, I think there are.

    I do not get the impression that this study is based on trickery and malice either. In fact, I’m suggesting it would be a mistake to assume so, especially given incredible weakness (and absence in some cases - like scientific evidence of any actual rise in autism prevalence) of supporting science cited by DeSoto and Hitlan.

    I was writing my long response while you posted that last one to me. I think I answered most of your questions in my other post.

    You did, thank you.

    I think where the overall argument runs into a stalemate is that the authors have limited their scope of discussion to the results of Ip et al data supporting Holmes et al Hypothesis. And that all comes down to interpretation of the statistics.

    I disagree. This is an imaginary stalemate. The hypothesis put forth by the Holmes et al. paper has a great big “IF” in front of it. Data cannot support such a hypothesis without addressing that “IF” or employing fallacious logic.

    DeSoto is pretty clear that they aren’t arguing that the Holmes et al Hypothesis is correct, only that the Ip et al data supports it.

    I disagree here as well. I think DeSoto and Hitlan demonstrated that they’ve probably accepted this hypothesis as some sort of postulate, uncritically - and without so much as a single mention of the big “IF”, or one iota of scientific support for that big “IF”. Let’s look at that important piece of the Holmes et al. conclusion one more time.

    If reduced overall mercury elimination is related to hair elimination, then autistic infants will retain significantly higher levels of mercury in tissue, including the brain, than normal infants. In light of the biological plausibility of mercury’s role in neurodevelopmental disorders, our study provides further insight into one possible mechanism by which early mercury exposures could increase the risk of autism.

    Emphases mine.

    That’s one gigantic completely unsupported “If” and a “look” at a “hypothetical” “might”.

    How was this treated by DeSoto and Hitlan?

    “(the first haircuts would reflect mercury excretion in utero and very early life)”

    “This means that the more severe autistic cases actually had less excretion of mercury.”

    “But among the autistic children, there was no linear relation between the mother’s mercury exposure and excretion of mercury in the hair.”

    Mercury is not known to be anything other (like excreted) than passively taken up from the circulating blood during the formation of the hair follicle.

    Hi Dr. DeSoto [waves]

58. Comment by Schwartz — 21 November, 2007 @ 10:27 pm

    Ms. Clark,

    You can choose to behave any way you wish. However, if you want your arguments to be held with any credibility, they should be based on evidence or some sort of verifiable logic. Acting unprofessionally is hardly punishment against someone doing the same.

    I am quite familiar with the attitude of some academics toward non-academics, and I find I receive exactly the same attitude from many (not all) of the scientists who blog.

    Now, if you had some proof that Dr. DeSoto ommited a conflict of interest in the study, then that is a valid issue. But I have not seen any evidence of that. Given the fact that the editor asked her to do the analysis, I doubt you’ll trace the funding to some ulterior motive, but I might be wrong. However, there is no evidence to support that.

    Her motive for writing the study doesn’t really matter when looking at the merits of the study itself, especially when all the facts are fully disclosed. If someone had designed a study, collected data, hired a statistician and published results, without making the full dataset public, then perhaps the movtive should be taken into consideration. But here? There is nothing to hide!

59. Comment by Schwartz — 21 November, 2007 @ 10:38 pm

    Do’C

    I disagree. This is an imaginary stalemate. The hypothesis put forth by the Holmes et al. paper has a great big “IF” in front of it. Data cannot support such a hypothesis without addressing that “IF” or employing fallacious logic.

    You make a valid point. I will make a point to read the study, and see if this is the thrust of the conclusion and discussion. Your argument appears sound, but sometimes the intent of a paper is difficult to discern from quoted snippets.

    I disagree here as well. I think DeSoto and Hitlan demonstrated that they’ve probably accepted this hypothesis as some sort of postulate, uncritically - and without so much as a single mention of the big “IF”, or one iota of scientific support for that big “IF”. Let’s look at that important piece of the Holmes et al. conclusion one more time.

    Same comment as above. Your point is well articulated, and I will have to read the semantics very carefully when I go through the study. But I certainly got that impression from Dr. Desoto’s response here.

    However, the study purchase will have to wait until tomorrow. I need a clear mind for this, and I got suckered by this discussion for the last two nights and the lack of sleep is catching up.

60. Comment by Schwartz — 21 November, 2007 @ 10:44 pm

    Prometheus,

    I am not an expert at statistics, but in my wandering on the internet, I have read some commentary complaining about the problems with using the generally accepted p=0.05 for statistical significance.

    If I recall, the complaint was that with small sample sizes, this number is not adequate and I think they said that the bar for statistical significance should be raised higher as a result. I’m assuming in this case that would render the results even more insignificant?

    But I clearly recall the conclusion of the study being that medical study results using a P=0.05 will results in a high rate of incorrect conclusions being drawn from the results.

    I wish I could remember more…

61. Comment by Schwartz — 21 November, 2007 @ 10:51 pm

    Ms Clark,

    Where did you get the impression that the authors didn’t like blogs? The only reference outside the greeting is:

    We think Blogs are great. Open forums read by all sorts of folks. However, we will not be posting again on this website, please keep informed.

    I think you may be a bit oversensitive to the post’s tone or to the PhD title. Her post had some light digs, but about equal to the original post. As for not coming back? That’s her perogative and I didn’t take it as an insult, and I’m not a PhD either. I suspect they’re pretty busy making a single rebuttal posts on a number of blogs.

62. Comment by Schwartz — 21 November, 2007 @ 11:08 pm

    Daedelus2u,

    I am saddly familiar with the acrimony and bad blood. However, that does not justify behaviour in kind IMO. I did not find the Author’s post insulting, or done in bad faith.

    I do find unsubstantiated accusations of fraud or conspiracy completely inappropriate.

    Bringing wakefield into the equation doesn’t really do anything for me. Right now, I look at the facts about him. He published a controversial study. He has a very controversial hypothesis which has not yet been supported by credible evidence. His original study was retracted by many of the authors. He is under trial in the UK by the GMC. However, I was not aware that he was under trial for fraudulently fabricating results? As for Mr. Deer, his allegations have not been independently verified to my knowledge. He certainly has an incentive to publish sensational stories.

    I will reserve my judgement of Wakefield’s character and actions for the results of the trial once they are published, and I can examine the actual evidence. Unfortunately, right now, I only have access to a single biased synopsis of the trial.

    As I commented to Ms. Clark, examining evidence closely based on prior behaviour is fully acceptable. But disregarding an argument, or even stating allegations based on one’s position on a topic, without any evidence is a pretty flawed argument. As I stated before, the motive is also not important when evaluating a fully published argument.

    It is perfectly justified in being skeptical and you may actually be right about the funding source — which I assume would be a violation of conflict of interest guidelines. However, it is a flawed to throw accusations around without any evidence.

    IMO, this is the equivalent of someone ignoring anything published by the CDC or the IOM because they have supported flawed studies in the past.

63. Comment by Ms. Clark — 22 November, 2007 @ 3:23 am

    Can someone clarify this for me? Didn’t DeSoto come up with the idea of reviewing the Ip data all on her own or maybe in conjunction with Hitlan?

    “In June 2004, the Journal of Child Neurologypublished an article by Ip et al1describing a study of mercury lev-els in hair and blood of children with autistic spectrum disorder compared with normally developing children. The authors reported no statistically significant differences between the mercury levels in 82 children with autistic spectrum disorder and 55 normally developing control children.

    On May 21, 2007, the Editorial Office of the Journal of Child Neurology received a communication from Dr Catherine DeSoto expressing concern about what appeared to be obvious inconsistencies in the data analysis in the Results section of that article.”

    The fact that the Geiers had handed on a copy of this paper while the ink was still drying on it gives me enough reason to suspect Dr. DeSoto. She just happens to have NO background in autism or heavy metals and just happens to have a background in testosterone and estrogen and psychology.

    And she just happens to cite utter garbage, the Holmes paper (which is hysterical) and the Bernard paper (which is whatever is more hysterical than hysterical), and she misrepresents a paper about a baby with mercury vapor exposure as showing that: “Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered. [4]”

    They were trying to figure out what to diagnose him, the above statement makes it sound like he got a diagnosis of idiopathic autism.

    This kid was worked up for Lesch-Nyhan disorder! She doesn’t mention that the kid had lots of symptoms of mercury poisoning but they didn’t figure that out at first. She wants her readers to think that it’s happened lots of times where a kid is frankly mercury poisoned and doctors are thrown off and diagnose it as autism.

    That paper is perfect for showing how mercury does not cause autism. As long as the child had this continual exposure to mercury he was in pain, he was weak, and he basically was a wreck. As soon as they removed the exposure the kid started to get healthy and he was not autistic! The child was massively exposed to mercury compared to the tiny tiny bit in a vaccine. Why is this so hard to understand? Mercury does not cause autism, obviously. Even if you want to postulate some sensitive kids who survive microscopic thimerosal doses without becoming autistic what about the ones who have been massively dosed with mercury and did not become autistic! There’s a baby who got a huge dose of thimerosal (specifically) it was so concentrated that the tissue where it was injected (on the leg) died. The baby didn’t become autistic.

    DeSoto might be innocent of having ulterior motives, but I still think it’s a totally fair question given her background, the superficial looking connection to the Geiers. And the history of the PSC for digging up expert witnesses of a certain kind.

    All she had to say was, “No, I don’t intend to be an expert witness.”

    Notice that she has not said this, though she may yet.

    Her paper still *stinks,* in my opinion, and shows her very bad grasp on toxicology, physiology and autism. it’s amazing that someone with a PhD would make these kinds of mistakes.

64. Comment by Joseph — 22 November, 2007 @ 7:58 am

    with small sample sizes, this number is not adequate

    The problem is that small sample sizes are not adequate. The convention itself is OK.

    If you had huge sample sizes, you might get significance at that point (I have little doubt you would.) But then you might see that the actual difference in mean levels is small.

    Statistical significance is one thing. It only tells you something about the likelihood that an actual difference was found. The extent of the difference is something else.

65. Comment by Joseph — 22 November, 2007 @ 8:04 am

    The child was massively exposed to mercury compared to the tiny tiny bit in a vaccine.

    For reference, a thermometer contains 0.5 to 1.5 grams of mercury. We’re talking about 40,000 times the mercury equivalent of a TCV.

66. Comment by daedalus2u — 22 November, 2007 @ 10:42 am

    Schwartz, there was testimony under oath in the Autism Omnibus trial by the person who did the PCR tests on the samples that all of the positive results were false positives and that he told Wakefield that the positive results were all false positives before the paper was published. That transcript is publicly available. Wakefield has never addressed those issues except to be silent about them.

    In the UK, libel is taken very seriously. If a charge of libel is sustained, the loser pays damages and court costs. If you are libeled, you get your legal fees to bring suit reimbursed. Wakefield sued Brian Deer for libel. Brian Deer was ready and willing to go to trial. Wakefield withdrew his lawsuit and paid Brian Deer all of Brian Deer’s legal fees up to that point. Brian Deer has a photocopy of the check on his website. The only defense against libel is if the statements are actually true. By withdrawing his lawsuit, Wakefield has conceded he would not prevail in the lawsuit because the statements Brian Deer made were actually true.

    In thinking more about Dr DeSoto’s paper and statements, she is not an expert in mercury physiology or in autism. She has no prior publications in those areas, her CV doesn’t list any activities in those areas. If she was approached by the anti-mercury lawyers, and they gave her the “mercury causes autism” files to read, and that was all she read, she would have a very distorted and one-sided (and wrong) view of what the science actually says about mercury and autism. That may be precisely the lawyers’ plan. The legal case isn’t about science, it is about getting the “win” for the client. Any tactic that is not “illegal” a “good lawyer” is supposed to use to get the “win”. It isn’t about “truth”, or “justice”, it is about money and “the win”. If the lawyers can trick their “expert” into giving testimony that helps their client that is up to the opponents to catch any errors in it. Provided the lawyers don’t actually coach the expert to lie (but non-lawyers are under no such restraint).

    As I think more about it, I suspect that Dr DeSoto won’t testify at trial. That would give the opponents the opportunity to cross examine her and allow her to catch her errors about mercury poisoning being misdiagnosed as autism. Similarly, the statements she made here that “Holmes may be wrong, the hair excretion idea may be incorrect.” is not something that the lawyers would want to hear coming out of the mouth of one of their “experts”.

    With out her testifying, it is simply a statement by a “scientist” in a “peer reviewed” journal. I think that is why all the “scientists” who have published on the “mercury causes autism” idea were not called to testify in the first trial (and won’t be called to testify in any trial). Then the unpublished raw data could be asked about. For example, the raw data in the “A Case-Control Study of Mercury Burden in Children with Autistic Spectrum Disorders” could be subjected to a “real” statistical analysis. When a data set consists of 221 data points with a range from 0 to 58, a mean of 4.06 +/- 8.69, the data set is highly skewed. The mean reflects a few extreme values, not the majority of cases.

    The appreciation that chelation with DMSA in the absence of actually heavy metal poisoning likely does cause brain damage all by itself is not well appreciated.

    http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pubmedid=17384765

    Does Dr DeSoto appreciate that her paper will be used to justify treating children with chelation in the absence of any actual medical indication? Treatment that will likely cause injury? I doubt it.

67. Comment by Schwartz — 22 November, 2007 @ 8:02 pm

    Daedalus2u,

    I know you read my debate on Autism Diva’s site about the Chadwick testimony. The trial was not about Dr. Wakefield, so we hardly saw any appropriate rebuttal from anyone representing him. The Chadwick testimony was brought forth to directly challenge the testimony of Kinsbourne regarding his use of Wakefields data. I can hardly consider it a proven fraud since he did not have a chance to defend himself.

    It is even more suspicious that Chadwick did not seem to be a witness in the GMC hearing. Why would the GMC not include a charge of fraud against Dr. Wakefield and bring him as a witness at a trial where he was appropriately represented and prepared to defend himself? The testimony I have seen to date is far from enough to convict him of fraud and given the fact he hasn’t been charged with it to my knowledge, it stands to reason there isn’t adequate proof.

    Settling a lawsuit proves absolutely nothing — I know this from experience. It certainly does NOT prove the correctness of the defendant’s statement. It just means that the plaintiff did not want to continue further with the lawsuit. For you to conclude anything else is pure speculation.

    From what I can tell, Wakefield wasn’t even the one funding the libel trial so I think your statement is misleading.

    Additionally, that lawsuit only dealt with the accusation that Wakefield took ~ 50K pounds and didn’t declare it as a conflict of interest. That is not at all related to the accusations you leveled, or other accusations that Brian Deer leveled, so your reference to that case does absolutely nothing to help your argument here.

    So I stand by my statements on Wakefield. I will wait for the evidence that is presented in the proper forum, with proper defence to see what the outcome is. That he was not even charged with fraud makes me question the real strength of the Chadwick testimony, especially since he didn’t appear to be a GMC witness and it certainly would appear to be completely appropriate and relevant in that trial.

    As for lawyers and legal cases, yes, the objective is to win. That is why Medical studies are published and funded the way they are — they are the defence for Pharmaceuticals, or evidence for plaintiffs. Why is your rant so one sided? Legal tactics are employed by good lawyers for both sides. I would expect nothing less.

    Does Dr DeSoto appreciate that her paper will be used to justify treating children with chelation in the absence of any actual medical indication? Treatment that will likely cause injury? I doubt it.

    Now I’m baffled. You’re accusing Dr DeSoto of being Naive?, unethical? This is because someone in the future might just use this paper as conclusive proof that that mercury causes autism, and then go and treat their child with Chelation that results in serious harm? And IF all of these things happen, it’s on her head?

    Just step back a bit, and think about your line of reasoning here (Ms. Clark’s term Hysterical comes to mind). I don’t think I need to point out the logical flaws.

    Dr. DeSoto’s paper points out an error in the first study, and then claims that the results in the Ip et al paper support the Holmes hypothesis. That is the essence of the study as I understand it. On paper, it is an attempt to adjust the conclusion of one part of a body of evidence that exists on the topic, nothing more.

    If you have more evidence please present it before you blame Dr. DeSoto for every future mistreated Autistic child in North America.

68. Comment by terrip — 22 November, 2007 @ 8:38 pm

    I am assuming that DeSoto does not really care about being accepted as a serious scientist by the scientific community. If she ever came up for tenure and this article was reviewed by any of the well-respected scientists in the autism community…well, she would not fare very well. She’d be dismissed on the bibliography alone. I’m sure she’s aware of this, so I suspect her motivations must lie elsewhere. Only she knows. But the motivation is certainly not to make a contribution to the scientific literature. And I’m sure she knows who these peoples are — all the people at the STAART centers, at NIMH, at the major autism research universities, such as UCSD, Illinois, MIT, Pittsburgh, North Carolina, etc.

69. Comment by daedalus2u — 22 November, 2007 @ 9:35 pm

    I think naive. That is my default when I see someone doing something that I consider this suspect. Mercury physiology isn’t her field, neither is autism. To “get up to speed”, takes months and months of time. A busy researcher doesn’t have months and months of time to do research outside her field. No one can get “up to speed” without reading at least a couple hundred papers in the field at a minimum. It really takes a lot more.

    My guess (and it is a guess and I admit quite speculative) is that only a select portion of the literature dealing with mercury and autism was fed to her and she came to a faulty understanding based on what she read. It would be easy to do that if you only read what the anti-mercury people subscribe to.

    Am I “hysterical”? I am concerned that children are being injured every day by useless chelation which is damaging their brains. I am concerned that the court may make an error and destroy the vaccine industry. Courts do make errors. Dow Corning was forced into bankruptcy based on the false premise that silicone caused immune system dysfunction. It doesn’t. The vaccine industry is a lot more important than silicone breast implants.

    Cases of scientific fraud are not adjudicated in courts of law. The proper forum is the scientific literature, not a court of law. If he had a reply, the Lancet would have published it. As far as I am concerned, Wakefield is guilty of scientific fraud. Credible scientific witnesses have testified under oath that he published false results. The results he published have not been replicated even though researchers with greater skill and expertise have tried numerous times. He has been silent with respect to these charges. I have read enough that I see his work as fraudulent and won’t ever rely on it. Whether a court finds him guilty of fraud, no reasonable or reputable scientist will rely on his work ever again. Real scientists only want to use results that are reliable. Unreliable results are worse than nothing, they waste your time and give you false ideas and make any derivative work you do unreliable. I appreciate that you are not a scientist, and so might not have the same perspective.

    People have died because of the MMR hysteria he whipped up. I see Wakefield the same as someone who claimed women were witches 400 years ago so he could sell torches to the villagers as they burned them to death.

    How many deaths do there have to be before “hysteria” is justified? How many IQ points to children have to lose from useless chelation before “hysteria” is justified?

70. Comment by Schwartz — 22 November, 2007 @ 10:35 pm

    Daedelus2u,

    You can’t seriously be proposing that people should stop studying the effects of chemicals on children because some doctor might chelate someone as a result? You also can’t seriously believe that this study is going to change the mind of a doctor who believes that chelation is going to help? I don’t think this study will make an ounce of difference to the doctors using chelation therapy.

    I also can’t believe you fear for the vaccine industry. They are protected by law in the US since the mid 2000’s. Additionally, comparing vaccine lawsuits to purely cosmetic breast implant surgery is a pretty big stretch, let alone the fact that vaccine trials are not heard by jury (unlike the silicon breast implants), so you are comparing apples and sheep.

    I would also like to know why you think that Wakefield is to blame for deaths? The logical connection of the two would be a huge stretch. There was already a suspicion of MMR long before he published the study. In fact, I believe the MMR uptake rates in the UK were already dropping before his study was published. Even if you could prove that he was soley responsible for a decline in uptake (which is impossible of course), you would have to prove that every death due to Measles would have been prevented by the vaccine. That is also something you can’t prove. So sensational claims that can never be substantiated are pretty pointless.

    A small number of children also die from vaccines, that’s why parents are compensated in the US for this. Are you proposing that we should never investigate why these children die to attempt to improve the safety of vaccines like MMR just in case they find something that causes uptake rates to drop? Especially when there is scarce credible safety data to start with? Your arguments seem to be quite exaggerated based on your strong opinion in the matter.

    Cases of scientific fraud are not adjudicated in courts of law. The proper forum is the scientific literature, not a court of law.

    Here you are clearly mistaken. The GMC trial in England is specifically dealing with his treatment of the science. He has been charged with professional misconduct, in addition to other things.

    In reporting a link between MMR and a regressive autistic state, as he did in the Lancet, he was dishonest, irresponsible and misleading.

    This certainly appears to address the science and the specific Lancet study. So why no appearance by Chadwick if he has incontrivertable truth of fabricated results? Unless he did appear in the trial, although I’ve seen no mention of it to date and the plaintiffs have rested to my knowledge.

71. Comment by Schwartz — 22 November, 2007 @ 10:38 pm

    I meant to quote the bottom section:

    Cases of scientific fraud are not adjudicated in courts of law. The proper forum is the scientific literature, not a court of law.

    Here you are clearly mistaken. The GMC trial in England is specifically dealing with his treatment of the science. He has been charged with professional misconduct, in addition to other things.

    In reporting a link between MMR and a regressive autistic state, as he did in the Lancet, he was dishonest, irresponsible and misleading.

    This certainly appears to address the science and the specific Lancet study. So why no appearance by Chadwick if he has incontrivertable truth of fabricated results? Unless he did appear in the trial, although I’ve seen no mention of it to date and the plaintiffs have rested to my knowledge.

72. Comment by Isles — 23 November, 2007 @ 6:51 am

    Schwartz, it is you who is delusional if you think DeSoto’s paper doesn’t give encouragement to those who believe autism is mercury poisoning and chelation is the cure. Where do you think the paper was first circulated? It was taken directly to the Evidence (whoops, Environment) of Harm newsgroup with an introduction by Mark Geier. The obvious effect is to tell those types that they’re on the right track and to give them more august-looking journal articles to cite in their materials.

    You are also wrong when you argue that the false claims about vaccines do not threaten the availability of vaccines in a manner analogous to what happened in the breast implant industry. Why do you think the vaccine injury compensation system was created? Manufacturers were contemplating stopping the production of vaccines entirely because even a completely groundless lawsuit is expensive to defend against and the standard childhood vaccines are not a very profitable product to begin with. Vaccine Court petitioners are free to take their cases into civil court after a set pendency period or after the verdict if they don’t like it, so your “apples and sheep” claim, beyond evoking the specter of Boyd Haley, holds no water.

    As to Wakefield and MMR, maybe you are suggesting that it’s OK for Wakefield to have put the fear of God into parents and caused children to get the measles as long as nobody died from it. It is not all right. And there is a difference between responsible scientists doing postlicensure safety studies (which is routinely done, by the way) and a researcher-for-hire waving around results that tort plaintiffs paid him to produce.

73. Comment by 666sigma — 23 November, 2007 @ 9:11 am

    Wow. I am blown away be the level of detail and discussion on this topic. I guess some people have a better understanding of statistics than I thought.

    One comment: no one would ever use a one-tail analysis to test this assumption even if you thought it was one directional. However, this really does not affect the result. One tail or two is only a factor of 2 (or 1/2 depending on how you look at it. 90% confidence at two tails equals 95% at one tail. However, if this is done to move from saying 80% confidence to 90%, you could argue about false perceptions are being created. IMO, this just says you need more data.

    Of course, most of the data on autism is a lot more subjective than anyone wants to admit.

74. Comment by Interverbal — 23 November, 2007 @ 9:22 am

    Hi Schwartz,

    Well now that I am rested and have had my turkey I would like to address some points.

    “I was misled by the wording you used to describe the errata acknowledgment by the original authors. It did sound like this paper was frivilous in pointing out an error that had already been dealt with.”

    It was not our intent to mislead anyone. I appreciate the chance to clarify.

    “While you may feel this way, the joint article published with your name effectively mocked the publication. It did not state a mere preference for peer-review.”

    I have been uncertain as to whether or not to accept your rebuke. The tone is certainly very blunt. In the end I have decided to defend our wording because our logic is valid. I reject your rebuke.

    “We have Interverbal who is clearly black and white when it comes to analysis and we have the authors who argue that the data from the Ip et all study shows support for Holme’s hypothesis contrary to the conclusion of the authors.”

    I don’t think this is an accurate statement. I really can’t think of any core area of study that is black and white, statistics included. I think that specific situations modify standards. I think you can do this, but I think you need a good reason. I asked Dr. DeSoto to provide one if she can. I will open that question up to anyone who cares to address it.

    “The points of contention being treatment of outliers, but mainly interpretation of significance in the results and the resulting conclusion.”

    Actually the treatment off outliers doesn’t seem to be in contention at all. Unless someone cares to create one. The only issue I had with the DeSoto & Hitlan’s treatment of outliers is that they attempted to use it to justify aspects of their reanalysis.

75. Comment by daedalus2u — 23 November, 2007 @ 10:58 am

    Schwartz, you are not a scientist, so your ignorance of how scientists deal with scientific fraud is understandable. Scientific fraud is dealt with by other scientists in the scientific literature not by legal proceedings. A charge of scientific fraud is extremely damaging to a scientist’s career. It can destroy a lifetimes worth of work. It is taken extremely seriously. For a credible charge of scientific fraud to be answered with silence is (to me, and I think to most scientists) an admission of guilt. That alone is enough evidence to ignore and to reject what ever that scientist has ever published as being unreliable. There is lots of unreliable work in the literature. Most of it is unreliable through simple honest error. A real scientist doesn’t want any of his work to have errors in it. A real scientist will correct errors. A real scientist will retract a paper that is in error, even if the error arose from an honest mistake. It is only a dishonest scientist that won’t respond to allegations of scientific fraud.

    There is no formal investigatory body charged with (and funded for) evaluating suggestions or claims of scientific fraud. Courts are simply not equipped to handle such things. Courts are about protecting rights of the accused until they have been found guilty “beyond a reasonable doubt”. Scientists are about protecting the scientific literature. Figuring out reality is really difficult. When people lie about it, (as Wakefield did) it makes it much harder to figure out. When there is a credible allegation of scientific fraud, and that charge goes unanswered, the work of that scientist becomes suspect and is no longer considered reliable by other scientists. That scientist’s work no longer meets the Frye standard or the Daubert standard, and can’t until the charges of scientific fraud have been dealt with.

    http://en.wikipedia.org/wiki/Frye_test

    http://en.wikipedia.org/wiki/Daubert_standard

    Responding to serious allegations of error and fraud are part and parcel of the scientific method. Not responding is a serious deviation which renders the work unreliable. By either standard Wakefield’s work is unreliable and hence of zero scientific value, unless and until he answers and satisfactorily refutes the credible charges of fraud. That is the current state of Wakefield’s work, it is presumed unreliable until he provides a satisfactory response. That is how scientific standards for fraud should be.

    You don’t seem to understand that the integrity of the scientific literature is more important to scientists than the career (or the rights) of any individual scientist. Rejecting suspect work that is actually correct only slows down scientific progress. If it is correct, eventually that correctness will be apparent. Accepting incorrect work leads to other incorrect work and wastes time. If you based your work on something incorrect, your effort is wasted and of zero value.

    Here is a case where students working for a professor caught her committing fraud. They reported it, at enormous personal cost. 3 quit school having spent 16 years pursuing degrees they will never get. The others had to start over. More than 20+ person-years lost. Who bore that cost? The students. They did the right thing, but paid a terrible price for it. That is the price of being a responsible scientist.

    http://www.sciencemag.org/cgi/content/summary/313/5791/1222

76. Comment by Do'C — 23 November, 2007 @ 3:44 pm

    All,

    Please limit discussion to the DeSoto & Hitlan paper, relevant science, and related citations. Speculation about possible motivations, suggestion of unsubstantiated connections to other things (ie: vaccine litigation, listservs, or other individuals), and wandering and commentary that is not relevant (ie: Wakefield, etc.) is off-topic and inappropriate.

    Thank you

77. Comment by daedalus2u — 23 November, 2007 @ 4:05 pm

    Sorry, it is easy for me to get carried away and off topic. But as far as mercury exposure now vs. in the past, I think there is pretty good evidence is was a lot higher 60 years ago.

    In looking at PubMed under “pink disease”, a number of older references have been added since I looked a few years ago. There is extremely compelling evidence that pink disease was actually mercury poisoning, and mercury poisoning from a lot of mercury. Not a small number of children died from pink disease, averaging nearly 60 per year for a decade.

    Link

    The death rate due to pink disease in Australia was about 30 per million children in the late 1940’s.

    The main cause seems to be teething powders that were from 16 to 33% calomel, HgCl (85% mercury). Mercury excretion in the urine of some of these children with pink disease was measured at 50 to 150 micrograms/100 mL. That is unprovoked by chelation.

    The most popular teething powder in Britain, Steedman’s was 26.3% calomel, and sold 7 million doses a year. Another manufacturer made 30 million. The usual dose of mercury in these powders was a grain of calomel (65,000 micrograms). In some places half the children were given at least one dose of mercury containing teething powders. That would be 55,000 micrograms mercury.

    Link

    I think the exposure of children to mercury via teething powders was enormously higher 60 years ago than it was at the height of thimerosal containing vaccines.

    Link

    Pink disease was mentioned in the first Autism Omnibus trial, but that trial was more about measles and there was no testimony about it. It is commonly described as a “hypersensitivity” to mercury (mostly because only 1 in 500 got it), but when children are given 150,000 micrograms of mercury and they develop symptoms of mercury poisoning, I don’t think it is necessary to invoke “hypersensitivity”.

    Link

78. Comment by Joseph — 23 November, 2007 @ 5:04 pm

    For anyone interested in what the error was all about, Ip et al. had originally reported that the mean mercury level in controls was 17.68 nmol/L, with P of 0.17. This changed to 14.68 nmol/L with a P of 0.06.

    There was also apparently an error in mean hair mercury levels. They originally were 2.26 and 2.07 ppm for autistics and controls, respectively. After the corrections, this changed to 1.98 and 1.92 respectively. The P did not change in the hair mercury analysis.

    I do buy the explanation that the error was the result of typos. But it’s unfortunate that scientists would be this sloppy, specially when it comes to a politically charged scientific debate such as this one. Let’s face it, they basically encouraged the conspiracists with this.

    In light of the corrected data, one thing that completely baffles me is DeSoto’s claim that the data support Holmes et al. The autistic kids actually had a bit more mercury in the hair than the controls, even though we can’t say that for sure as there’s no statistical significance for that difference.

    Her reasoning is that since the blood level difference was “almost” statistically significant, but the hair level difference was far from being statistically significant (even though it was in the same direction) then this somehow supports Holmes et al.

79. Comment by Schwartz — 23 November, 2007 @ 8:10 pm

    Interverbal,

    I have been uncertain as to whether or not to accept your rebuke. The tone is certainly very blunt. In the end I have decided to defend our wording because our logic is valid. I reject your rebuke.

    Rebuke is a stronger word than I would have intended, but still accurate as a re-read my response. I do think it justified Dr DeSotos defensive response so we’ll have to differ on this point.

    I don’t think this is an accurate statement. I really can’t think of any core area of study that is black and white, statistics included. I think that specific situations modify standards. I think you can do this, but I think you need a good reason. I asked Dr. DeSoto to provide one if she can. I will open that question up to anyone who cares to address it.

    Prometheus’ posts provided more insight into this since my post and I am satisfied with your answer. If one wants to apply a different standard, then there should be some specific justification.

    Actually the treatment off outliers doesn’t seem to be in contention at all. Unless someone cares to create one. The only issue I had with the DeSoto & Hitlan’s treatment of outliers is that they attempted to use it to justify aspects of their reanalysis.

    OK. I didn’t get that nuance on my first read of your article.

    Well now that I am rested and have had my turkey I would like to address some points.

    I’m glad you had a good turkey day. It was a quiet day up here in Canada.

80. Comment by Schwartz — 23 November, 2007 @ 8:15 pm

    All,

    Please limit discussion to the DeSoto & Hitlan paper, relevant science, and related citations. Speculation about possible motivations, suggestion of unsubstantiated connections to other things (ie: vaccine litigation, listservs, or other individuals), and wandering and commentary that is not relevant (ie: Wakefield, etc.) is off-topic and inappropriate.

    Thank you

    Do’C, I am biting my tongue to respond to accusations of ignorance and delusion, but this is your house.

81. Comment by Do'C — 23 November, 2007 @ 8:41 pm

    Schwartz,

    I hear you. I’m sure that does not help, but I do appreciate your willingness to stay on-topic to DeSoto and Hitlan.

    Luck of the draw for my timing in reading the most recent comments (after a most decent Thanksgiving Day off myself) did not appear to favor you.

82. Comment by Joseph — 24 November, 2007 @ 6:38 am

    Another item of interest… looks like Dan Rossignol has also misrepresented Chrysochoou et al. in a presentation titled “Evidence of Metal Toxicity in Autism” which you can find on the web. A slide shows the paper, with the caption “Mercury Poisoning Mistaken For Autism.”

    http://tinyurl.com/2ac5v9

    Perhaps DeSoto read this presentation by Rossignol before she was aware of Chrysochoou et al. ?

83. Comment by daedalus2u — 25 November, 2007 @ 10:28 am

    Schwartz, I meant no malice toward you in any of my comments. My malice is reserved for those with demonstrated ill intent (simple error does not warrant malice to me).

    Given the mean and the standard deviation of the two data sets (autism and non-autism), what are the expected values of the upper and lower values at various confidence limits? In other words, the important “number” is not a difference in the average values, but rather how much overlap is there in the two distributions.

    Since an “autism diagnosis” is done on an individual, what does an individual mercury level tell about membership in which group? In other words, if we take an individuals blood mercury level and see where in the distribution if falls, does it improve an ability to diagnose autism? There may be a difference in the average, but if the two distributions overlap a huge amount, that difference doesn’t mean anything. If an elevated blood level of mercury is somehow indicative of autism, what level implies what autism likelihood?

    In the Faroe Islands, the geometric mean is 8.82 micrograms/L or 44 nanomoles/L.

    http://aje.oxfordjournals.org/cgi/reprint/150/3/301

    What would this study “predict” the incidence of autism in the Faroe Islands would be? More than 75%?

    If we are going to use blood mercury as a “diagnostic” test, what are the implied false negative and false positive rates depending on what value is used as the cut-off? If the difference in the mean is due to a few extreme cases, the false negative and false positive rates might be so high any test would have no value.

    Because mercury physiology involves thiols which are highly involved in oxidative stress physiology, it would not be a great surprise if there were differences in mercury physiology between individuals depending on their state of oxidative stress. ASDs do tend to be in a state of higher oxidative stress. Everything that is perturbed by oxidative stress will be observed to be perturbed in ASDs, including steroid physiology, energy physiology, the immune system, just about every physiological system is perturbed by oxidative stress, it is a very powerful modifier of physiology (which is why physiology uses it as a control parameter). But these are all effects of the state of oxidative stress, not causes. A modest difference in mercury physiology would be consistent with the data in Holmes et al (not with their hypothesis), and even with the “dose-response” effect that they saw, that the most severe ASDs (i.e. most severe oxidative stress) had the most perturbed mercury physiology.

84. Comment by Do'C — 25 November, 2007 @ 10:33 am

    daedalus2u wrote:

    A real scientist doesn’t want any of his work to have errors in it. A real scientist will correct errors. A real scientist will retract a paper that is in error, even if the error arose from an honest mistake.

    I agree with this sentiment. I think it remains possible that Dr. DeSoto may as well.

    Interestingly, Dr. DeSoto’s faculty webpage at UNI shows reference to a critical thinking page.

    http://www.uni.edu/desoto/ThinkCrticially.htm

    Additionally, I noticed that she is listed as supporting the rejection of “attempts to represent Intelligent Design as a scientific endeavor”.

    http://faculty.cns.uni.edu/~demastes/UNI_statement.htm

    Perhaps she is the type of scientist who will consider retraction/correction of what I think most geniune scientifically minded and critical thinkers would consider to be glaring mistakes for a peer-reviewed publication.

    Potentially:

    The implication that actual autism prevalence is increasing - without providing scientific support.

    The implication (and statement for the record, here) that Chrysochoou et al. represents a case where the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.

    The statement that mercury poisoning has “sometimes” been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered - without providing anything close to adequate scientific support (multiple supporting citations).

    The statement, “but the observable symptoms of acute mercury poisoning have been reported to match up with many of the problems observed in autism” - without referencing more recent, peer-reviewed literature.

    The statements that mercury is “excreted” in the hair - without providing scientific support.

    The implication that their data supports the Holmes et al. hypothesis without addressing that big “IF” regarding hair elimination that prefaces that hypothesis.

85. Comment by Joseph — 26 November, 2007 @ 5:53 pm

    The implication that their data supports the Holmes et al. hypothesis without addressing that big “IF” regarding hair elimination that prefaces that hypothesis.

    And again, on the surface the data seems to contradict Holmes et al. Unless they try to wiggle out of it like James Adams, who said the Holmes et al. hypothesis might only apply to very young autistic children.

86. Comment by Schwartz — 26 November, 2007 @ 9:56 pm

    Daedelus2u,

    No malice taken from your post.

87. Comment by Ableize disability information — 7 December, 2007 @ 2:31 pm

    Hey you have some really good stuff here, been reading for over an hour so far

    p.s any chance of a web link exchange?

88. Comment by RH — 7 January, 2008 @ 1:35 pm

    The DeSoto & Hitlan (2007) article has been featured on dozens of blogsites. We appreciate that our research has garnered so much attention, but have become increasingly aware that a great amount of misunderstanding exists and much misinformation is being repeated. For example, one blog site has stated that we did not do analyses which we clearly did do, and made it appear that the only problem with the Ip 2004 article was that a one-tailed test should have been used. We are troubled that this information about what we wrote is being repeated, since it is not accurate on several counts. It is our sincere hope that interested parties will try to keep an open mind and carefully read our article for what it says — and what it does not say.

    Please click on the link below for answers to Frequent Misunderstandings:

    http://www.uni.edu/desoto/desoto_hitlan_autism.html

89. Comment by C DeSoto — 7 January, 2008 @ 3:54 pm

    I hope you will allow your readers the benefit of this post.

    http://www.uni.edu/desoto/desoto_hitlan_autism.html

90. Comment by Matt — 7 January, 2008 @ 11:43 pm

    Dr. Desoto,

    I noticed a comment on your FAQ implying that you beleive this blog regularly refers to those it disagrees with as idiots.

    I guess “believe” and regularly are somewhat ill defined, but my google search of this site didn’t pick up that many hits.

    The one on this page is a reply, not a statement by the owner–most of the eight hits on the google search (note that there are over 150 blog posts here) are replies or quotes of others.

    Was there a reason you felt that you needed to include such an untested belief in your FAQ?

    Matt

91. Comment by Interverbal — 8 January, 2008 @ 12:02 am

    Drs DeSoto & Hitlan,

    Thank you for visiting this blog. It is rare to have authors visit blogs to discuss their work and I am quite appreciative.

    I have read your FAQ very carefully. It does indeed clarify several points. Your FAQ also made it clear that several points in contention are so grey area that they may not be worth the time right now.

    However, there are several important issues that your FAQ does not address. Do’C specifically listed these several posts ago:

    1) The implication that actual autism prevalence is increasing - without providing scientific support.

    2) The statement that Chrysochoou et al. represents a case where the preliminary diagnosis was “autism of unknown etiology” later shown to be due to mercury exposure.

    3) The statement that mercury poisoning has “sometimes” been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered - without providing scientific support (multiple supporting citations).

    4) The statement, “but the observable symptoms of acute mercury poisoning have been reported to match up with many of the problems observed in autism” - without referencing more recent, peer-reviewed literature.

    These issues are not misrepresentations, nor are they answered in your FAQ.

    You can choose whether or not addressing these issues is worth your attention.

    Thank you both very much for your time.

92. Comment by Connor — 8 January, 2008 @ 8:41 am

    I had been reading all this before the holidays and checked back today. Before I had believed most of the postings on this site were valuable and useful commentary, but I don’t think that anymore. You slam a researcher for possible minor mistakes and completely misrepresent and write blatant false information about what they said. Your post definitely gives the idea that the only problem with the original article by Ip was they should have used a one tailed test. Moreover, you stated they did not do tests they did do. I plan to write to the authors and thank them for clarifying things for me. But I am writing to you to say I am disappointed that you have taken so much time and effort to confuse the issue. You should take your misinforming posts down if you want to be seen as legitimate source of information. Clearly the “tale of two tails” is a fairy tail written to “confuse the issue”.

    Me think you protesteth too much. … the only explanation I can imagine is that you are trying to hide by the smoke and mirrors? if so what?
    http://www.youtube.com/watch?v=Rn5-VN3SH1o

93. Comment by Do'C — 8 January, 2008 @ 7:40 pm

    It is our sincere hope that interested parties will try to keep an open mind and carefully read our article for what it says

    You don’t mean things like:

    “Furthermore, mercury poisoning has sometimes been presumptively diagnosed as autism of unknown etiology until the mercury poisoning has been uncovered. [4]“

    And

    “This means that the more severe autistic cases actually had less excretion of mercury.”

    — and what it does not say.

    Such as appropriate supporting scientific evidence for the above.

    How about evidence for that “increase in autism”? Not diagnoses, autism iself (as written)?

94. Comment by Interverbal — 8 January, 2008 @ 8:20 pm

    Hello Conner,

    You write “You slam a researcher for possible minor mistakes and completely misrepresent and write blatant false information about what they said.”

    We did make several mistakes which were clearly acknowledged.

    You write “Your post definitely gives the idea that the only problem with the original article by Ip was they should have used a one tailed test.”

    Then I am afraid you did not read our post very carefully.
    You write “Moreover, you stated they did not do tests they did do.”

    The issue here was whether a test of the corrected data was run using a one-tailed test with the outliers kept in. The authors never mention that they ran such a test and I incorrectly stated that the authors did not run this test. This issue arose in the comments, it was not part of the actual critique. I have acknowledged this error multiple times now and in several different locations.

    “But I am writing to you to say I am disappointed that you have taken so much time and effort to confuse the issue.”
    We sometimes make errors too. When it is clear to us that we made an error, we take steps to acknowledge and/or fix it. However the issues we were wrong about is hardly the whole of our critique. I recommend that you re-read our critique as well as the most recent discussion to get a feel for what the talking points are.

    Also, neither Do’C nor myself seeks to “confuse the issue”. If you feel that we in fact are attempting to confuse the issue or do a smoke and mirrors job, then you probably will want to find a another blog to participate on that you feel is more ethical.

    Furthermore, several important issues were not resolved or even discussed by the authors here, or in the FAQ page. They have listed multiple times in the recent comments. To point these out as problems hardly counts as misinformation. In consideration of this I see no reason to accept your rebuke.

    Thank you for your time.

RSS feed for comments on this post.

Important Comment Notes

Installed spam filter plugins may result in your comment being held in a moderation queue. Please consider making a copy of your comment "just in case". You can use many XHTML tags to mark up your comment if desired.

Additional information about comments and other contact can be found here

Leave A Comment

Name or Pseudonym (required)

E-mail (required, not published)

URL (optional)