Tracking the Cognitive Profile of Autism Over Time: A Landmark Analysis of Intellectual Disability Trends in the ADDM Data

New peer-reviewed article finds key signature of thimerosal in the trends.

Cynthia Nevison’s 2025 paper in Science, Public Health Policy & the Law, published today, presents a novel and methodologically careful analysis of the cognitive severity of autism spectrum disorder (ASD) over time using the CDC’s Autism and Developmental Disabilities Monitoring (ADDM) Network data. While ASD prevalence has been widely reported to be rising, this study is the first to systematically track trends in the co-occurrence of intellectual disability (ID, IQ < 70) in ASD diagnoses from birth years 1992 to 2014.

The key metric examined is the “ID fraction”—the proportion of ASD cases with IQ below 70—as an indicator of autism severity. Nevison found that the ID fraction declined from 48% in birth year 1992 to a minimum of 31% in 2002. It then remained relatively stable through 2006 before steadily increasing again, reaching 40–41% by 2014. This trend contradicts the prevailing assumption that rising ASD prevalence is largely attributable to expanded diagnostic criteria capturing more mildly affected, higher-functioning individuals.

Nevison’s analysis includes multiple cross-checks for confounding. Three different weighting methods were used to calculate national averages, accounting for differences in data availability and state participation across ADDM reporting years. The influence of changing racial and ethnic demographics was also assessed. While these factors explained a modest portion of the trend, they were insufficient to account for the abrupt drop in ID fraction from 2000 to 2002 or its subsequent rise after 2006.

Importantly, these inflection points align with changes in vaccine policy regarding thimerosal, a mercury-based preservative. The sharp drop in the ID fraction coincided with the phase-out of thimerosal from most pediatric vaccines around 2001. Conversely, the increase in the ID fraction after 2006 coincided with the CDC’s expanded recommendations for flu shots—including thimerosal-containing multi-dose formulations—for both infants and pregnant women.

The study also highlights a crossover in ASD prevalence by race and ethnicity: between 2008 and 2010, Black and Hispanic children began to show higher ASD prevalence rates than White children, a trend that continued through 2014. These groups also had consistently higher ID fractions. By 2014, nearly 2% of Black children in the ADDM sample were diagnosed with ASD and co-occurring ID.

Nevison proposes that renewed prenatal and early-life exposure to thimerosal and other immunologic stressors (e.g., aluminum-containing Tdap vaccine during pregnancy) may play a role in modulating ASD severity, particularly among low-income populations with higher public health insurance coverage. These hypotheses are consistent with a broader literature on toxicant-induced neurodevelopmental injury but have not been rigorously explored in current CDC analyses.

The study is careful to distinguish between hypothesis generation and causal inference. While the timing of policy and ID fraction changes are consistent with a thimerosal effect, Nevison does not claim that thimerosal is the sole or primary driver of ASD prevalence. Rather, she suggests it may influence the distribution of severity within the diagnosed population.

In conclusion, this study contributes a previously missing dimension to autism epidemiology: a consistent, long-term measure of cognitive severity. It also raises testable hypotheses about environmental modulation of ASD risk and severity. Future analyses will need to independently validate these patterns and further investigate the intersection of prenatal exposures, demographic trends, and ASD outcomes.

Nevison C. Time Trends in United States Autism Prevalence with Co-Occurring Intellectual Disability: Is There a Signature of Thimerosal? Science, Public Health Policy and the Law. 2025 Jun 25; v7.2019-2025