Yet Another Thimerosal Study

In what could be considered an early Christmas gift, researchers in Poland have published yet another study that fails to find an association between thimerosal and autism. Well, I say Christmas gift, but really it is the sort of like the bunny suit that Aunt Clara made for Ralphie in the Christmas Story movie.  After you open the box and look at it you are left wondering what was she thinking.

If you don't know what all the drama is about when it comes to thimerosal in vaccines and autism, well you must have been hiding under a rock for the past decade.  You might want to just go back there and skip this post, it will be better for your sanity.

Anyways, I can't talk about the details of the study because it is in Polish and I can't read Polish.  But two things did jump out at me as I was reading the abstract.

First, vaccines that contain thimerosal are still in routine use in Poland.

That surprised me a little bit since the safety of using thimerosal as a preservative in vaccines was first questioned more than twenty years ago. I thought that most of the first and second world countries had moved away from using it by now.

Second, you can't just a study by its title.

The title of this study is "Lack of association between thimerosal-containing vaccines and autism" and most of the abstract matches the title.  They included children with and without autism, determined how much thimerosal each child was exposed to (not sure how), adjusted for other confounding factors, yada yada yada, and concluded "no significant association was found between TCVs exposure and autism."  Just what you would expect from your run of the mill thimerosal study.

But then there is this one little sentence -

After adjusting to potential confounders, odds ratios of the risk of autism developing for infants vaccinated with TCVs were 1.52 (95% CI: 0.29-11.11) for doses 12.5-87.5 microg, 2.78 (95% CI: 0.29-11.11) for 100-137.5 microg and 1.97 (95% CI: 0.37-18.95) for these exposed > or = 150 microg

I had to read that twice before it sank it.  Here the abstract just finished saying there was no significant association and then it says that infants exposed to thimerosal had an increased risk of autism?  And this risk went up (more or less) as the dose of thimerosal increased?

Say what?

Of course the answer to this conundrum rests in the word "significant", as in statistically significant.  The researchers could not be confident that the relationship was not due to random chance.  And the 95% confidence intervals listed reflect that fact.

For example, if you take look at that first set up numbers, a child was almost one and a half times more likely to develop autism if they were exposed to anywhere between 12.5 to 87.5 micrograms of thimerosal than a child who was not.  But there is a 95% chance than the odds could have been anywhere from one third less likely all the way to more than 11 times more likely.

Or in other words, you can be mostly confident that you have no clue what the real relationship is.

Even still, I would love to be able to read this study and see what the numbers were.  Because it looks like there might be something interesting going on with the data.  Is the data evenly distributed and that is causing the wide intervals or is the data grouped together, possibly highlighting a vulnerable subset?

If only I could read Polish.

Anyway, even if this study did show an increased risk of autism after exposure to thimerosal and that the risk increased as the exposure to thimerosal increased, it would not be anywhere near enough to overturn the other studies that show the opposite.  All it could do is put a little doubt in your mind that something was missed in the earlier studies.


References

Mrozek-Budzyn D, Majewska R, Kieltyka A, Augustyniak M. [Lack of association between thimerosal-containing vaccines and autism]. Przegl Epidemiol. 2011;65(3):491-5. Polish. PubMed PMID: 22184954.