Maternal hair samples were taken from all mothers of the 61 thimerosal-exposed and control infants.  Breastmilk samples were taken from 8 mothers (no information given on whose mothers). 

A simplistic mathematical model was used to estimate the half-life of ethylmercury based on the distribution of blood mercury levels in 21 blood samples taken on different days from 21 different children.

B.  What did the authors find?

Note:  In order to compare results, some conversions are required.  To convert micrograms (mcg) into nanograms (ng), multiply by 1000 (1 mcg = 1000 ng).  To convert nanograms into micrograms, divide by 1000.  The molecular weight of mercury Hg 200.6.   Therefore 1 nmol Hg = 200.6 ng Hg.  To convert nanomoles (nmol) into nanograms, multiply by 200.6.

The authors set statistical significance at p = 0.05.  Anything less than or equal to a 5% probability that the data is due to chance is judged to be significant.

1.  In the exposure group, vaccines used had the following amounts of mercury per dose: DPT 25 mcg, Hep B 12.5 mcg, and Hib 25 mcg.  The mean mercury doses received were:

Two month olds:    45.6 mcg (range 37.5 - 62.5 mcg)   
Six month olds:   111.3 mcg  (range 87.5 - 175.0 mcg) 
(It is unknown if this difference is statistically significant.)

2.  Out of the 21 blood samples from thimerosal-exposed children that fell within the "range of quantitation," the average mercury concentrations were as follows:

Two month olds:  8.20 nmol/L (SD=4.85)   or    1.645 mcg/L
Six month olds:    5.15 nmol/L (SD=1.20)   or    1.033 mcg/L
(The levels for 6 month olds were not significantly lower,  p = 0.06.)

3.   The only data for the control group within the "range of quantitation" came from one blood sample from a two month old child.  It was 4.90 nmol/L or 0.983 mcg/L.   There were significantly more samples that fell within the range of quantitation for the exposure group (21 out of 33) than the control group (1 out of 15) (p=0.04).

4.  Mercury was detected in urine samples in only 4 out of 27 thimerosal-exposed children, and none of 14 control children.  (It is unknown if this difference is statistically significant.)   

5.  For the 4 urine samples with detectable mercury from the exposure group, the values were 3.8 nmol/L for a two month old, and 5.75 nmol/L (SD=1.05) for 3 six month olds.   The values in mcg are 0.762 mcg/L and 1.153 mcg/L.  (It is unknown if this difference is statistically significant.)   

6.  Out of the 22 stool samples from thimerosal-exposed children, the average mercury concentration was 81.8 ng/g dry weight (SD=40.3) for two month olds, and 58.3 ng/g dry weight (SD=21.2) for six month olds.  (The level for 6 month olds were not signifcantly lower, p = 0.098.)

7.  The average mercury concentration in stool samples taken from the 9 children in the second control group was 22 ng/g (SD=16).   These 9 children had significantly lower levels of mercury in their stool than the exposure group  (p=0.002). 

8.  Maternal hair values were 0.45 mcg/g for the exposure group and 0.32 mcg/g for the control group.  Mercury concentrations in maternal hair were not significantly different between the exposure group and the control group.

9.  Breastmilk from 8 unidentified mothers averaged 0.30 mcg/g, ranging between 0.24-0.42 mcg/g of mercury.

C.  What were the flaws of the study?

1.   Invalid sampling of the study population.    There is an overwhelming number of questionable exclusions in an already small sample size.  

a.  The authors claimed to have studied 61 children. The small initial sample size of this study means any conclusion, even if all 61 children were used to provide data, would have to be interpreted with extreme caution. 

In reality, they studied data from only 21 blood samples, 31 stool samples, and 27 urine samples from separate children.   There is no information on how many of these samples were from the same children. 

b.  There is no information on why stool and urine samples were not taken from all subjects.  When exclusions are not objectively justified, the assumption that the sampled data gathered is representative of the whole population becomes invalid.  If the sampling is invalid, the stool and urine data gathered cannot be generalized to other children.  This means the stool and urine results cannot represent the study population as a group.

c.  Although they did try to take blood samples from all 61 children, they excluded almost 2/3rds of those blood samples for two dubious and poorly defined reasons.  The first reason was insufficient blood volume, defined as less than 1 mL.   They lost 17.5% (7 out of 40) of the exposure samples and 28.6% (6 out of 21) of the control samples because they failed to draw enough blood for the study.  The loss of  such a large percentage of subjects from failure to obtain adequate samples is not valid and scientific sampling.

The second was mercury levels falling outside the "range of  reliable quantitation."  The authors said that that mercury levels found below the range of 2.5 to 7.5 nmol/L, depending on blood volume, were excluded.  However, they showed a data plot (Figure 1 on pg 1739) where 17 out of 21 other samples with levels within that range were included.  The only factor that can explain why there were included was blood volume of these samples, which was not defined.  There is no explanation on why some samples that fell within the same range were excluded, while others were included.

Furthermore, there is no information on the detection limits, or what minimum amount of mercury must exist for the instrument to detect the mercury.  It would be very different if all samples excluded fell below the detection limit.  This would have been an objective criterion.  But the samples had sufficient mercury to be detected, so detection limit was not the reason for exclusion. 

Unjustified exclusions means the blood concentration results cannot represent the study population as a group.  For example, it was very nice that the 21 children whose samples were accepted fell below a safe level of mercury, but they cannot be judged to be representative of all children who received vaccines containing thimerosal in the study, let alone all such children outside of the study.

2.  Unknown mercury exposure.    Without knowing how much of the mercury exposure came from vaccines, no conclusions can be drawn about the effects of vaccines on mercury levels measured.

a.  No samples were taken from the subjects before vaccination.  This would have been the easiest and cheapest way to control for other sources of mercury exposure.  Without taking a comparison sample, how could the authors have known if blood concentrations of mercury were "raised" after vaccination, let alone determine how much they were raised and that any amount raised was safe?

Omitting a pre-vaccination sample from the design effectively rendered any connection with the subject of vaccines indefensible.   Any conclusion that mentioned vaccines, such as "thiomersal in routine vaccines poses very little risk to full-term infants," is completely unsupported by the information available in this study.  

Cold vapor atomic absorption does not distinguish whether the mercury found is ethyl or methyl.  If they had used another method which could have identified which kind of mercury is found, that would have sufficed as well. 

b.  It is unknown how much mercury the children were exposed to from other sources such as diet and air pollution. 

The authors tried to address this problem by studying mercury levels in a control group.   However, the only control data obtained came from ONE blood sample in the entire study and 9 stool samples from children who were not formally part of the study.   Because the rest of the blood and urine samples were badly confounded by inadequately justified selections and exclusions, any difference between the exposure and control group cannot be accepted as scientifically valid.

The authors also tried taking maternal hair and breastmilk samples.  Those samples showed that, indeed, the mothers have been exposed to mercury.  But it is unknown if maternal mercury levels came from mercury exposure that is shared by their children (e.g. if their exposure came from their vaccines or drugs or came from past residence in a polluted area).    Even if one assumes that the mothers' mercury exposure is shared by their children, we still do not know how much.  Other studies sample maternal hair during pregnancy, when shared exposure with the fetus is known and can correlate with cord blood and other measurements after birth.  In this study, maternal hair was quite irrelevant.  Breastmilk mercury levels would have been useful, had they been taken from all breastfeeding mothers and correlated with their children's mercury levels.

3.  Ridiculous calculations of  half-life.

The authors admitted that this was not a formal pharmokinetic study, where the same subject would be tested repeatedly for mercury at regular intervals after exposure.  However, they felt that taking single samples from 21 children at "various time points after exposure" could suggest a "half-life of less than 10 days."  In other words, they could estimate the half-life by pretending all 21 samples were from the same person.  This assumption is ludicrous, especially given the extremely small sample size used.  Science is the study of observed data, not imaginary data or imaginary relationships between the data.

The mathematical model they used for calculation of half-life was simple, but it did require at least two samples from the same subject to be able to calculate any change in that subject.   Taking two samples from two different children and pretending their difference represents a change in the same child is a dishonest misuse of the mathematical model used to measure rate of change.

4.  Unconscionable conclusions by the authors.

a.  "Administration of vaccines containing thiomersal does not seem to raise blood concentrations of mercury above safe values in infants." 

(1).  They have no scientific data in this study to support any statement about the effects of vaccines on increased blood levels.  As explained above, since only one sample per child was taken, there is no basis to determine that there was even an increase, let alone attribute an "increase" to vaccines. 

(2).  Because of invalid sampling, blood concentrations measured cannot be generalized to other children in this study, let alone children at large.  

(3).  The "safe values" refer to the reference dose established for methlymercury.  No such values have been established for ethylmercury found in thimerosal.  While researchers often compare the two, it is duplicitous to generalize the safety of ethylmercury by standards set for methylmercury, without any qualifications.

b.  "However, no children had a concentration of blood mercury exceeding 29 nmol/L (parts per billion), which is the concentration thought to be safe in cord blood; (18) this value was set at ten times below the lower 95% CI limit of the minimal cord blood concentration associated with an increase in the prevalence of abnormal scores on cognitive function tests in children. "  [no reference provided.]

(1).  The reference (footnoted #18) provided for 29 nmol/L as" the concentration thought to be safe in cord blood," is titled:

18.  Nielsen JB, Andersen O, Grandjean P.  Evaluation of mercury in hair, blood, and muscle as biomarkers for methylmercury exposure in male and female mice. Arch Toxicol 1994; 68: 317-21.

Nielsen's paper is about mice and has no information that be construed as defining safe levels of mercury in human cord blood.  The second half of the sentence did not even have a reference.   Without references, the authors fail to present any evidence whatsoever in their paper to justify the safe levels of mercury used for comparison.  I have written Dr. Pichichero, the principal author, on February 20, 2003 asking for correct references, and will update this review if correct references are ever given.

Although the authors do not provide a reference, the National Academy of Sciences publication, Toxicologic Effects of Mercury, specifies a "BMDL [Benchmark Lower Dose] of 58 ppb of Hg in cord blood (corresponding to a BMDL of 12 ppm of Hg in hair)."  (Footnote#1)  This dose is equivalent to 289.1 nmol/L of methylmercury in cord blood.  This dose is then divided by the EPA recommended Uncertainty Factor of 10 (to allow room for error), to equal 29 nmol/L of methylmercury in cord blood.

(I would like to thank Sallie Bernard, Executive Director of Safeminds.org, for helping me find this reference.)

(2).  Even though this value is considered "safe," many experts caution that there may be no truly safe level with methylmercury.  Dr. Kathryn Mahaffey of the US Environmental Protection Agency says, "This is not a 'no observed adverse affect level.'  Within the Faroese cohort's data, effects at exposures less than those associated with a maternal hair mercury concentration of less than 10 ppm have been reported raising questions about whether or not a threshold for methylmercury's effects exists."  (Footnote#2)  It is unknown if ethylmercury is more or less harmful than methylmercury and what the true safe levels are, if there are any.

(3).  The authors speak of "blood mercury" as if methylmercury and ethylmercury were equivalent in effect and had the same safety levels.  Again, it is disingenuous to not to point out the limitations of this comparison.

c.  "Ethylmercury seems to be eliminated from blood rapidly via the stools after parenteral administration of thiomersal in vaccines."

"We estimated the half-life of mercury in blood after vaccination to be 7 days...."

"Assessment of these samples suggested that the blood half-life of ethylmercury in infants might differ from the 40-50 day half-life of methylmercury (range 20-70 days) in adults and breastfeeding infants.  The concentrations of blood mercury 2-3 weeks after vaccination noted in our study were not consistent with such a long half-life, but suggested of a half-life of less than 10 days."

(1).  Elimination means a decrease in mercury levels obtained at two different times from the same child.  Since only one blood sample was taken per child, there is no basis for determining there was a decrease in anything. 

(2).  Because they did not take a pre-vaccination sample and no immediate post-vaccination sample, there was absolutely no evidence in this study to conclude that mercury concentrations in blood came from thimerosal in vaccines.   Mercury was excreted in the stools, but there is no information that the stool mercury came from the vaccines.  For all the evidence presented in this study, vaccine mercury could have traveled through the blood in the first day to be stored in brain tissue, to never be eliminated or to be eliminated very slowly over years, and the mercury found in blood and stools were incidental remnants of dietary mercury exposure.  This study has no data to prove or disprove this hypothesis anymore than it has data to prove the conclusion that vaccine mercury is eliminated quickly. 

(3) Half-life is a function of change within the same subject.  Any calculation of half-life based on the difference in 21 concentrations taken from 21 different children is deceptive.  It is based on an illusion, a magic trick, that people will automatically assume that each individual child's results can stand in for results that 20 other children would have gotten on the same day.  Science is supposed to tease out and reject this kind of spurious reasoning.  Instead, the authors packaged it and tried to sell it to the public.

Let me illustrate.  Suppose a salesman came to your door, selling a pill called Mercury-Gone.  He tells you that taking one pill of Mercury-Gone will eliminate the total mercury in your blood by half within 10 days.  He says this claim is supported by a clinical study of 21 children who took Mercury-Gone.  They measured blood levels of Adam on Day 1, Ben on Day 2, Charlie on Day 3, Dan on Day 4, and so forth.  They found that John on Day 10 had half the mercury levels as Adam on Day 1.  They calculate that Mercury-Gone appears to reduce mercury levels by half in 10 days.  What would people call such a study?  Junk science?  Quackery, perhaps?

D.  What can be concluded from this study?

1.  Mothers of infants in this study were exposed to mercury, source unknown.  There was no significant difference between the hair mercury levels in mothers of subjects in the exposure and the control group.  Since the maternal hair samples were not taken during pregnancy, their relevance to the study is tangential at best.

2.  Mercury is more prevalent in the stool samples than in the urine samples of selected infants in this study.  But this prevalence could have been an artifact of subjective selection.   Therefore, this result has no scientific meaning.

3.  Stool concentrations of mercury were significantly higher in 22 exposure subjects than in 9 subjects used as controls, but who were not formally part of the study.  Though, this might be interpreted that the source of the mercury in stools was vaccines, this comparison is so confounded that such interpretation is not scientifically defensible.  We do not know how the exposure and control subjects were selected, and why others were excluded.  Sampling has to be rigorously objective, or anyone can obtain any result they want simply by picking their subjects carefully.  Therefore, this result has no scientific meaning either.

4.  Twenty-one children who had received vaccines containing ethylmercury had blood concentrations of mercury below the safe levels established for methylmercury.  It is unknown which type of mercury was in their blood and where this mercury came from.  Because of invalid sampling procedures, their results cannot be generalized to other children.

Because the paper fails to provide objective evidence meeting minimal scientific standards (or even normal common sense), no other conclusion is justified. 

References:

1.  National Academy of Sciences.  Toxicologic effects of methylmercury.  Washington, DC:  National Research Council, 2000.  (Page 12) http://www.nap.edu/books/0309071402/html/

2.  Dr. Kathryn Mahaffey. Development of Methylmercury Reference Dose. US Environmental Protection Agency.  http://www.masgc.org/mercury/abs-mahaffey.html

------------------------------------
The author would like to thank the Vaccine Science discussion list for their input and feedback.
http://groups.yahoo.com/group/VaccineScience/

Permission is granted to forward or reprint this article on the condition that it is reproduced in its entirety without any changes (everything between and including the 2 asterisk lines).   If this article is reprinted on another website, the author would appreciate a note with the link to the other website.  Her email address is list@freedom2think.com.

*****************************************************************