Autism spectrum disorders, endocrine disrupting compounds, and heavy metals in amniotic fluid: a case-control study.

The sample was small, and the authors stated that statistical power was expected to be low. Many exposure analyses used only samples with detectable levels, so the number analysed for chemicals was smaller than the final group. The study tested many chemicals and biomarkers, and most individual associations did not survive correction for multiple comparisons. Cases had more congenital malformations than controls, which may affect the comparison even after adjustment or exclusion. Amniotic fluid samples came from pregnancies selected for amniocentesis, not a random general-population sample. Diagnoses came from registers, and no clinical autism symptom data were available for validation. Historic stored samples and sample processing may have removed or underestimated some PFAS. The principal component finding combined several chemicals and receptor measures, so it cannot identify which exposure was responsible. The study cannot show that prenatal exposure caused autism.

The work was supported by Aarhus University, Statens Serum Institute, and the Danish Research and Innovation Council. The supplied text does not report individual author competing interests.

Do not read this as evidence that prenatal PFAS or heavy metals increase autism risk. Most associations were null, and the few signals that survived correction pointed the other way.