Is DMSA safe and effective?
November 26th, 2009
Yes!
For lead poisoning.
And probably as a treatment for mercury and cadmium poisoning (but not for assessing the body burden of mercury).
And possibly as a treatment for arsenic and antimony poisoning.
But as a treatment for autism, it hasn’t been shown to be either effective or safe, despite the titles of two articles (or one two-part article) in BMC Clinical Pharmacology.
These two articles, published by a diverse group of authors - including Professor of Materials Science and Engineering James B. Adams, PhD (lead author), DAN! practitioner Jeff Bradstreet, MD and Professor of Pediatrics and Section Chief of Pediatric Allergy, Immunology and Rheumatology, Jane El-Dahr, MD - are titled:
Strangely enough, given that the lead author is on the faculty of Arizona State University and the “anchor” author is on the faculty of Tulane University, the study was done under the auspices (and approved by the IRB of) Southwest College of Naturopathic Medicine - which Dr. Adams lists as his academic affiliation in these articles.
Curious.
It’s almost as though they couldn’t get a real university IRB to approve their study. After all, Tulane University has an IRB, and so does Arizona State University, but they decided to go with the IRB at a naturopathic college, instead.
Again, that’s curious.
But that’s not the last curious thing about these studies. In fact, the next curiousity about these studies is the very…..eccentric nature of the study design.
Let me start by outlining the odd and circuitous path of their enquiry into the safety and efficacy of DMSA as a treatment for autism.
Perhaps it is the writing style or maybe it is the chaotic nature of their research showing through in the articles, but it is very hard to track exactly how many subjects were in each of the several “phases” and sub-phases of this study. Here is what I could extract from their writing:
The study began with 82 autistic children, ages 3 - 8 years, who underwent a physical examination, blood tests for kidney and liver function, blood counts, and red blood cell (RBC) glutathione. If the examination and blood tests (with the exception of RBC glutathione) were within normal range (the article uses the phrase “not below the normal range” - curious), they were “eleigible to participate in Phase 1″.
Each parent filled out an initial Autism Treatment Evaluation Checklist (ATEC - a test devised by Drs. Rimland and Edelson and never properly validated) and a “Heavy Metal Exposure Questionnaire”. The parents also collected a “baseline first-morning urine sample” - for “heavy metal testing”. (three guesses which lab did the testing)
Here’s where it gets weird: the next step was to randomly divide the subjects into two groups. One group got a glutathione-containing skin lotion (to be applied once a day for a week) and the other group got a placebo lotion. At this point, I found myself asking, “What’s up with the lotion? Who’s selling the glutathione lotion?“, because this meant that - at best - their results would be much harder to interpret.
Further complicating the study, they had every subject receive a single “round” of oral DMSA (10 mg/kg/dose, three doses a day for three days - nine doses in total) to test their “heavy metal excretion” (and here I thought that autistic children were “poor excretors”). Only those subjects who had post-chelation urine “toxic metals” (as defined by “Doctor’s Data Laboratories” - aluminium, antimony, arsenic, beryllium, bismuth, cadmium, lead, mercury, nickel, platinum, thallium, thorium, tin and tungsten) greater than the 95th percentile (for people who had not received a chelating agent) continued on to “Phase 2″. (see: Mercurial Laboratories for a more in-depth discussion)
Amazingly, eight (8) children had all of their “toxic metals” below the 95th percentile even after chelation. This deserves to be a published result on its own! In all, 17 subjects failed to complete “Phase 1″ (testing, glutathione lotion and one “round” of DMSA), leaving only 65 (79%) to move on to “Phase 2″.
In “Phase 2″, the subjects who had recieved the glutathione lotion were scheduled to receive oral DMSA and those who had received the “placebo” lotion received a placebo capsule. This was to go on for an additional 6 “rounds”, each “round” consisting of three days of 30 mg/kg DMSA (divided into three doses) followed by eleven days of no treatment. The article is a bit confusing about this, saying that the subjects took the DMSA or placebo for “…up to 3 rounds” in one place and a few paragraphs later saying that “When the participants finished Phase 2 (after either 3 or 6 rounds)…” The reason for this is that urine testing was done after the second “round” of DMSA in Phase 2 (third “round” total) and those who were not continuing to excrete elevated levels of “toxic metals” after three “rounds” of DMSA were considered to have completed Phase 2. Why some had an additional “round” after this testing is unclear.
Before entering Phase 2, the 65 remaining subjects underwent further blood tests and also had the ADOS administered. The ADOS testing found that 7% of those who completed Phase 2 did not meet the criteria for Autistic Spectrum Disorder - we are not told how many of those entering Phase 2 failed to meet criteria. The authors did not seem particularly bothered by the fact that 7% of their subjects were apparently not autistic or on the “autistic spectrum” - they explained this away by stating that “All children continued on in the study, since they all had a previous clinical diagnosis of ASD.” [emphasis added]
In addition to the ADOS, the parents filled out the Pervasive Developmental Disorders - Behavior Inventory (PDD-BI - a screening test evaluated only by its originators) and the Severity of Autism Scale (SAS - a test developed by the authors and not validated).
Following completion of Phase 2, the 41 remaining subjects (50% of starting subjects) underwent repeat blood testing and an ADOS evaluation. The parents filled out the ATEC, PDD-BI and SAS questionnaires again, as well as a Parental Global Impression (PGI - a new test not yet validated) questionnaire.
Let’s recap the study so far:
The study consisted of three arms:
[1] Subjects who received placebo lotion and one “round” of DMSA - 15 subjects.
[2] Subjects who received glutathione lotion and three or four “rounds” of DMSA. It appears that five subjects were in this group - it remains unclear why some had three and some had four.
[3] Subjects who recieved glutathione lotion and seven “rounds” of DMSA - 21 subjects.
The testing consisted of RBC glutathione, complete blood counts, liver function tests, renal function tests, and urinary “toxic” and “essential” metal excretion as well as the ATEC (an un-validated test), PDD-BI (a screening test), the SAS (an unvalidated test), the PGI (new and unvalidated) and the ADOS (validated for autism diagnosis, but not severity rating).
The problems with this study are legion, but their results were the most damning part of the whole mess, because the results don’t say what the authors think they do.
Behavioral Testing:
In the behavioral testing, only the ADOS (which - as I mentioned before - is not valid for comparing autism severity) showed any diffrerence between the “got one round of DMSA and placebo lotion” and “got three to seven rounds of DMSA and glutathione lotion” groups. The other tests showed no statistically significant difference. Of course, the authors try to spin that as best they can, but the results speak for themselves.
Safety Testing:
DMSA’s safety was studied long before than this study, and its side effects are well known: reduction of white cell count (generally reversible), reduction in platelet count (generally reversible), liver injury (generally reversible) and some subtle decrease in intellectual function when given to children (and rats) with low (or zero) lead levels (not reversible).
The “safety testing” in these articles, strangely enough, did not include measures of the most significant (and the only irreversible) side effect that has been noted with DMSA, namely cognitive/intellectual functions (see: Dietrich et al and Stangle et al). Thus, in a supreme moment of irony, the authors have done exactly what some of them have long claimed that “the government” has done - failed to research the correct mode of toxicity.
The authors (and many, many other practitioners) are giving DMSA to children with low levels of heavy metals - exactly the group found to be at risk for permanent intellectual impairment. Yet they do no intellectual testing. They could have gone to Dietrich et al and read it right out of their paper, yet they didn’t.
Curious.
Biochemical Effects:
This area, which the authors call “Medical Effects” has the most bizarre results ever. Rather than spend the next year discussing them all, I’ll give my “highlights”:
[1] “Toxic metal” excretion.
Not surprisingly, lead excretion was up significantly - this is what DMSA does. However, in a major upset for the autism-is-mercury-poisoning hard-core, mercury excretion was fifth, after tin (?), bismuth (??) and uranium (????). Who knew that autistic children were so heavily contaminated with uranium? There was even more thallium excreted than mercury, over the inital “round” of DMSA administration.
While lead is still ubiquitous in urban areas - a legacy of decades of tetraethyl lead in gasoline, where are these children getting their exposure to tin, bismuth and uranium? Or is this yet another problem with the laboratory? The authors make a brave stab at explaining how their results differ from an earlier Bradstreet “study”, but it doesn’t really work too well.
[2] “Essential metal” excretion:
Several studies have looked at “essential mineral” excretion with DMSA administration - most have found that DMSA increases the excretion of zinc and copper to a minor degree. However, Adams et al found that potassium and chromium were the most significant losses during DMSA treatment.
Part of this comes from their…..eccentric way of looking at “essential mineral” losses as a percent of the RDA. This artificially elevates the “significance” of micronutrients like chromium (children in this age range should have about 15 micrograms a day).
The major problem with their analysis of this section is that they appear to have no idea how these elements get into the urine. Potassium, for example, is found in almost all foods and the urinary levels will fluctuate depending on how much there is in the diet and how much the body needs. DMSA has no effect on an alkali metal ion like potassium and the authors’ discussion makes it appear that their knowledge of basic physiology and biochemistry is inadequate.
[3] RBC glutathione:
This is a bit harder to interpret for a number of reasons. First off, the mean RBC glutathione doesn’t change significantly (it goes down in those treated with DMSA, but apparently not significantly). The baseline of all 72 subjects who had RBC glutathione done was a mean of 501 (+/- 246) - those 38 (this is confusing - which 38 are they referring to? Why not 41?) who went on to receive a second measurement in “Phase 2″ had a baseline mean of 523 (+/- 280). Two months after one “round” of DMSA, the ”38″ remaining had a mean of 478 (+/- 83). Another graph shows that those with higher RBC glutathione ended up with lower RBC glutathione after treatment and vice versa.
Secondly, their laboratory (Immunosciences) uses an odd “reference range for adults” - 427 - 714 micromolar (I assume it is micromolar - the article never gives the units - curious). There are a number of references (e.g. Richie et al 1996) that looked at large numbers of “normal” adults and found the range to be higher (670 - 1600 micromolar in whole blood; 1600 - 2800 micromolar in RBC’s).
There was also the problem of determining how the glutathione was measured. The authors state that Immunosciences Lab used an Oxis Research kit in which “…the absorbance measured at 405nm is directly proportional to the GSH concentration.” It may be nit-picking, but currently available Oxis Research glutathione measurement kits use absorbance at 400 nm, 420 nm and 412 nm - none use absorbance at 405 nm. Maybe it was an old kit.
Still, the change in the standard deviation of the RBC glutathione is curious (there’s that word again), but not terribly significant unless they can explain how reducing half of the subjects’ RBC glutathione is a “good thing”. Their presentation of the data raises more questions than it answers, such as “Why don’t you have RBC gluathione for all 41 subjects who completed Phase 2?” and many, many others.
Another point that the authors didn’t address about the RBC glutathione tests is that they don’t seem to show any positive effect of the glutathione lotion. Of course, they don’t break it out for us to show differences (if any) between the glutathione lotion and placebo lotion, but I suspect they would have if there had been any indication that the glutathione lotion had an effect.
[4] Platelets:
According to my university’s hospital lab, the “normal range” for platelets in children 3 - 6 years old is 204,000 - 402,000 per microliter of blood. This study showed that the baseline platelet count of their subjects who completed the study (n=41; I wish they would explain why the numbers keep bouncing around - I suspect it is because they didn’t get a full set of data on all of their subjects) had a mean (a rather useless value - median would be so much better) of 388,000 per microliter (+/- 274,000). Rather than give us a before and after number, they only give the percent above an below the “reference range” after one “round” of DMSA treatment. It’s almost as though they’re trying to hide something.
Figure 5 gives away part of the secret: like the RBC glutathione levels, the platelet counts show a reduction in standard deviation without a significant change in mean (or median - from my reading of the chart). While this may be an “interesting” (or “curious”) finding, it will take a lot more work to show that it means anything.
Bottom line:
These studies show neither efficacy nor safety of DMSA in the treatment of autism. This is not surprising, given the poor design of the study.
In fact, if I were to try to design a study that couldn’t show any results, I would be hard-pressed to do better than this one. The fact that the authors are so convinced that the DMSA did work is a testament to their pre-conceived notions.
In the end, this study gathered a bunch of data and then threw it against the wall to see what stuck.
And not much - if any - did stick.
Considering how long we’ve been hearing that DMSA treatment is dramatically ”curing” or “recovering” autistic children, the results from this study are distinctly underwhelming. It shows that there is no clinically significant difference between a single “round” of DMSA and multiple “rounds”, which suggests that DMSA doesn’t work at all. The authors’ assessment that a single dose of DMSA “did the trick” is a pathetic post hoc attempt at spinning the results in favor of their preferred outcome.
About the only thing it shows for certain is that glutatione lotion doesn’t increase RBC glutathione. But they didn’t address the efficacy of the glutathione lotion at all.
Curious.
Prometheus
Filed under: Autism Practitioners, Autism Science, Autism Treatments | 27 Comments »