Pages Menu
Categories Menu

Posted by on Mar 24, 2009 in Ethics, Neuropsychiatry

Tell Me, Is the Capital of France Still Paris?

Tell Me, Is the Capital of France Still Paris?


In an effort to initiate some kind of corrective, but arguably foolhardy, dialog with the editors of the Journal of the Neurological Scienceswho recently published an autism-biomarker study by Geier et al (see critical posts here and here, for example)I submitted a Letter to the Editor that identified (what I believe to be) serious flaws in the published work. My letter was published online by the journal March 9 (DOI: 10.1016/j.jns.2009.02.309 [subscription required]). The response letter submitted by Geier et al was published online March 19 (DOI: 10.1016/j.jns.2009.02.305). 

Without reproducing the letters, which would violate copyright, I can describe my published criticisms of the JNS article by Geier et al and the authors’ responseswhich, in my opinion, are compelling only in their creativity.

Criticism 1: Geier et al presented a cherry-picked introduction of published information that causally connects environmental exposure to mercury with autism, while implying support for this link from the Centers for Disease Control and Prevention.

But in their introductory discussion, Geier et al neglected to acknowledge the known pharmacokinetic differences between environmental mercury (primarily methylmercury) and ethylmercury; the growing series of epidemiologic studies that have found no association between ethylmercury-containing vaccines and the risk of autism; and authoritative statements made by the CDC, the Institute of Medicine, and the World Health Organization that deny any persuasive evidence for an autism-mercury link.

Instead of acknowledging these facts, however, Geier et al respond that my comments imply “only isolated statements by researchers associating mercury poisoning with autism.” This is a different argument altogether. I don’t contend that there’s limited published information associating mercury exposure with autism. There’s a lot of published crap out there contending all sorts of unfounded nonsense. Certainly Geier et al themselves have contributed liberally to the medical literature that attempts to link mercury exposure with autism and, unfortunately, they have found sympathetic publishers.

My point is that Geier et al refuse to acknowledge the facts raised and, specifically, the series of epidemiologic studies that negates an association between mercury exposure (whether by methylmercury or ethylmercury) and autism. Geier et al completely sidestep this criticism; but moreover, they go on to make a flabbergasting analogy between a mercury-autism relationship and a smoking-lung cancer relationship.

Criticism 2: Metabolite data (specifically for plasma cysteine, plasma oxidized glutathione [GSSG], and plasma total sulfate) of Geier et al are substantially different from values in the literature, including articles cited by the authors themselves.

In their response, Geier et al note that their values for plasma reduced glutathione (GSH) were consistent with those in the published literature. So their initial defense boils down to, “We got 1 out of 4 right.” Next Geier et al note that they measured circulating free cysteine, not total cysteinedespite the fact that they cited Han et al as their assay reference, a study providing plasma cysteine values approximately one order of magnitude higher (160-360 µmol/L) than their values (eg, 17.8 µmol/L).* Geier et al go on to cite Mansoor et al (an article not mentioned in their JNS study) as a reference for their cysteine values. In this case, Mansoor et al found a mean plasma reduced cysteine value in their male controls of 14.5 µmol/L, which, in fact, is closer to the cysteine values of Geier et al.

However, Mansoor et al imply that free (ie, non-protein-bound) cysteine exists in both reduced and oxidized forms. So in this case, their mean control value for free cysteine in men would be 14.5 + 88.2, or 102.7 µmol/L. This summed value is consistent with values in other studies, which indicate that the free cysteine level in human plasma is approximately 50% of the total cysteine value and therefore remains substantially higher than the values obtained by Geier et alfor example, 112 µmol/L ± 15.2 in 13 volunteers aged 24-29 years (Guttormsen et al) and 140 µmol/L ± 21 in 10 healthy fasting subjects aged 31-52 years (Suliman et al).

Geier et al then thank me for “raising a question about the units for our [GSSG] measurement.” But they claim they had prior knowledge (before submission of my letter to the JNS editor) of a typographical error and state that the units should have been µmol/L, not nmol/L.

Let me just say that this claim isHow can I put it?very interesting. As documented at this blog, I e-mailed Henry Sershen, associate editor of Neurochemical Research on October 14, alerting him to the discrepant metabolite values of Geier et al, including those for plasma GSSG (which were similarly published in JNSsee Criticism 3):

The associate editor of Neurochemical Research, Henry Sershen, forwarded my e-mail to Mark Geier (anchor author), who reviewed the information and responded to Dr. Sershen that the published values for GSSG were consistent with those of James et al in a 2004 study (which was also cited by Geier et al). In fact, a review of this study reveals that the printed mean value for plasma GSSG in control children, 0.32 nmol/L ± 0.1 (range, 0.11-0.43 nmol/L), is very similar to the control value of Geier et al (0.35 nmol/L ± 0.05). But it is also very different from the control value published by James et al in their 2006 study, 0.24 µmol/L ± 0.1, or 240 nmol/L ± 100.

To that end, I e-mailed Dr. James asking for clarification of the plasma GSSG values in her articles, and she responded that the units in the 2004 article were incorrect and should have been µmol/L, not nmol/L. Consequently this correction negates Dr. Geier’s rebuttal that his values for plasma GSSG are consistent with those of James et al.

So after defending his original GSSG units and citing James et al as a confirmatory reference (which was acknowledged to be incorrect by the lead author), Dr. Geier (as the anchor author of Geier et al) now claims that the GSSG unit notationmade in both the Neurochemical Research and JNS articleswas a typographical error.

Last, it should be noted that Geier et al, in their letter response, do not address their discrepant values for total plasma sulfate.

Criticism 3: The possibility of significant overlap between subjects in the JNS study and those in a nearly simultaneously published study by Geier et al in Neurochemical Research is raised. If overlap existed in the 2 study populations, it should have been acknowledged by the authors.

To this point, Geier et al flat out deny that their JNS data were previously published (although, they do not specifically address any overlap of subjects between the 2 studies). This claim is made despite the fact that their control metabolite values (including standard deviations) are identical between the 2 publications, and the fact that their metabolite values in Neurochemical Research (n = 38) and JNS (n = 28) are very similar.

Plasma Measurement

Mean ± SD

ASD Subjects of Geier et al
(n = 28)

ASD Subjects of Geier et al
(n = 38)

Control Values of Geier et al

Cysteine, µmol/L

17.8 ± 9.5

17.8 ± 8.3

23.2 ± 4.2


0.46 ± 0.16

0.48 ± 0.16

0.35 ± 0.05

Total sulfate, µmol/g protein

924 ± 245

934 ± 252

1930 ± 184

Criticism 4: Geier et al failed to control for age-related differences in urinary porphyrin excretion, given that their subjects ranged in age from 2 to 16 years (for a rationale, see Minder and Schneider-Yin).

Geier et al rebut that I did not read “Table 1” from their study, which compared urinary porphyrins with similar mean and age ranges. They claim, therefore, that subject age should not have affected their results to any significant extent. However, Geier et al fail to note that the age ranges of these groups were still very wideeg, 3-13 years. 

Geier et al then write that 3 “independent” studies from 3 continents “have now demonstrated that porphyrinuria is concomitant with [autism spectrum disorder].” One of these studies is, not surprisingly, from Mark and David Geier, who obtained their porphyrin data from the Laboratoire Philippe Auguste. The author of another autism-porphyrin study is, in fact, at this very laboratory. 

* In a previous post at this blog, the possible distinction between total cysteine values and free cysteine in the work of Geier et al was addressed.

bmartin (1127 Posts)

A native East Tennessean, Barbara Martin is a formerly practicing, board-certified neurologist who received her BS (psychology, summa cum laude) and MD from Duke University before completing her postgraduate training (internship, residency, fellowship) at the Hospital of the University of Pennsylvania in Philadelphia. She has worked in academia, private practice, medical publishing, drug market research, and continuing medical education (CME). For the last 3 years, she has worked in a freelance capacity as a medical writer, analyst, and consultant. Follow Dr. Barbara Martin on and Twitter.