Print Friendly, PDF & Email
April 20, 2014 | 2 Comments


  1. There are a lot of allegations in this article, some of which are very serious. I’m surprised to see them presented here without supporting documentation.

    The SSRIs have been around for over 25 years now and have been taken by a hundred million or so people by this point. In spite of this extraordinarily large base of human epidemiological data, no consistent pattern has emerged showing a clear connection to developmental disorders and autism. In contrast to the statement in the 4th paragraph of this article, there are multiple papers in the peer-reviewed literature which have failed to find any connection between SSRI use in pregnancy and adverse development outcomes, including Hviid (2013), Sorensen (2013), Austin (2013), and Udechuku (2010), who reported that 6 of 7 identified studies found no long term developmental effects of exposure during pregnancy. Other papers have found an increased risk of autism/development disorders associated with SSRI use in pregnancy, in most cases the effect size was small and it was not possible to confidently reject the alternative hypothesis of confounding by indication.

    Clearly medication use during pregnancy should be limited to cases of extreme need, whether one is talking about SSRIs or other common medications having a much clearer track record of adverse pregnancy outcomes. This being said, it is unclear to me exactly what sort of animal studies done in the 1980s would have made abundantly clear a risk that is remains hotly debated after 25 years of use in hundreds of millions of people. The evidence certainly does not seem to be there to an extent that justifies the allegations in this article. I’ve worked for many years with scientists performing preclinical safety studies. I’ve never met one who would be more likely to cover up an adverse teratogenicity study than the author would be to recommend suicide to a patient. Nor has support been offered here for the attack on the motivations of the Motherrisk group other than that they published conclusions that the author disagreed with.

    The data in medical science is often difficult to interpret due to the many vagaries of confounding, covariance, and the ever present problem of population heterogeneity. Interpretation is always further complicated by bias and conflict of interest, and at times by outright dishonesty. Nonetheless, it seems civilized to assume the former until convincing evidence for the latter becomes clearly evident. Otherwise personal attacks too quickly become a shortcut around the much more difficult and productive process of analyzing the data.

  2. This response is filled with argument that is not relevant to the issue. The fact that 100 million people have taken SSRI’s is beside the point with respect to whether they cause autism. This is straight out of the Pharma defense handbook. What is useful to know is how many women who were pregnant and on SSRI’s in the relevant trimester that the brain develops. The animal studies done by the companies themselves demonstrated teratogenicity and it is in some of the labels (Lexapro/Celexa/Paxil) but not all. They all demonstrated the ultimate birth defect (death-one of Wilsons four principles of teratogenicity) but not all labels say so. Further there have been many animal studies that have been looking at this issue since the late 1980’s and almost all come to the same conclusion (Jean Lauder/Margaret Kirby/Shuey/Sadler among others. In addition almost every epi study done without ties to Pharma has found increases risk of birth defects and SSRI’s. The ones that haven’t were designed to fail and did fail. They also have deep Pharma ties (Margulis and the new Lee Cohen paper)
    Gideon Koren (Motherrisk) has had some very public ethical issues related to drug company studies.

Leave a Reply