Prenatal Exposure to Perfluorinated Compounds (PFC) and Autism Spectrum Disorders: Results From the Early Markers for Autism Study

Background:  Perfluorinated Compounds (PFC’s) are chemically stable, persistent chemicals widely used in industrial applications as well as consumer products. Several have been identified as developmental and immune system toxicants. Greater than 98% of the US population had detectable serum levels of several classes of PFC’s in a NHANES survey (2003-2004). However, no studies have examined the potential link between prenatal PFC exposure and risk of autism.

Objectives: To determine the association between PFC concentrations in maternal blood samples collected during mid-pregnancy and risk of autism spectrum disorders (ASD’s) in the child.

Methods: The study population came from the Early Markers for Autism (EMA) study, a population-based nested case-control study of children born from 2000 to 2003 in Southern California. Children diagnosed with an ASD (N = 430) or a developmental delay but not autism (DD, N = 328) were identified from the Department of Developmental Services (DDS). Controls (N=439) were randomly sampled from birth certificate files and matched to ASD cases on gender, birth month, and birth year.  ASD and DD diagnoses were validated based on review of DDS medical records.  Following expert review, 136 children were reclassified from the DD group to the ASD group (final ASD N=566).

All mothers were participants in California’s prenatal expanded alpha-fetoprotein screening program. Maternal serum samples were collected at 15-19 weeks gestational age.  After routine prenatal screening testing was complete, left-over samples were stored frozen and later analyzed for concentrations of 8 PFC’s (PFOA, PFOS, PFNA, PFHxS, PFDEA, PFOSA, Et-PFOSA, and Me-PFOSA) using a high-performance liquid chromatography-tandem mass spectrometry device and an on-line solid phase extraction method. Exposure levels in the controls were used to develop quartile cutoff points.  Unconditional logistic regression was used to model the odds ratios comparing prenatal exposure in cases vs. controls. Regression models were adjusted for demographic variables as well as matching variables.

Results: PFC median concentrations (ng/mL) in this population were lower than those reported in the NHANES populations (PFOS Medians: NHANES 1999-2000 = 30.2, NHANES 2003-2004 = 21.2, EMA = 17.90; PFOA Medians: NHANES 1999-2000 = 5.1, NHANES 2003-2004 = 4.1, EMA = 3.60).  PFDeA and PFOSA concentrations were mainly below limits of detection, consistent with NHANES results (PFDeA = 83.3% <LOD, PFOSA = 75.4% <LOD).  Preliminary analyses indicated that PFOA concentrations were negatively correlated with odds of developing ASD (OR, upper quartile vs. lower quartile: 0.64, 95% CI [0.42, 0.97], p-value = 0.033). Risk of ASD was not significantly associated with prenatal exposure to other PFC’s, though further analysis is necessary. 

Conclusions: Preliminary results indicate that concentrations of PFC’s during pregnancy in the EMA population were similar to, though somewhat lower than, the NHANES population. Concentrations of PFC’s varied widely, both by demographic and diagnostic groups. Exploration of main effects as well as differing effects of PFC’s by subgroups of demographic variables will be further explored in this large, diverse population.