While individuals diagnosed with autism spectrum disorders (ASD) have higher levels of antibodies directed towards gliadin, a component of wheat gluten, no study has examined anti-gliadin antibodies (AGA) in etiologically relevant periods before diagnosis. The objective of this study was to investigate if maternal levels of AGA, during pregnancy and at the time of birth, are associated with ASD in offspring. We analyzed AGA in archived neonatal dried blood spots (NDBS) for 921 ASD cases and 1090 controls, and in paired maternal sera collected earlier in pregnancy for a subset of 547 cases and 428 controls. We examined associations with ASD diagnoses as a group and considering common comorbidities (intellectual disability [ID] and attention-deficit/hyperactivity disorder). We compared 206 cases to their unaffected siblings to examine the potential for confounding by shared familial factors. Odds of ASD tended to be lower among those with the highest levels (≥90th percentile) of AGA compared to those with low levels (<80th percentile; OR 0.78, 95% CI 0.56-1.09, measured in NDBS). This pattern was more apparent for ASD with comorbid ID when measured in NDBS (0.51, 0.30-0.87), with a similar trend in maternal sera (0.55, 0.24-1.29). High levels of AGA were similarly associated with lower odds of ASD in the sibling comparison. In summary, we found little association between maternal antibodies raised against components of gluten and risk of ASD in general. Exposure to high levels of AGA in the pre- and perinatal periods may be protective in terms of risk for ASD with ID. LAY SUMMARY: There is a debate among both scientists and community members as to whether an immune reaction to gluten exposure could be considered a cause of autism. We examined antibodies that are directed against gliadin, a part of gluten, in samples collected from pregnant mothers and their newborn babies. We did not see any major differences in the antibody level among those children diagnosed with ASD or their mothers compared to children who were not diagnosed with ASD. High levels of the antibodies were in fact associated with a somewhat lower risk of ASD with co-occurring intellectual disabilities, though we cannot tell from this study why that might be the case.