Summary: A large Danish registry study of more than 1.1 million births found little evidence that most maternal health conditions during pregnancy directly cause autism in children. Instead, nearly all previously reported associations between maternal diagnoses and autism appear to be explained by familial factors—genetic inheritance and shared environmental influences—rather than a direct causal effect of the mother’s medical conditions.
After rigorous analysis that included comparisons within families and evaluation of fathers’ health records, only complications that directly affected the fetus remained strongly linked to autism. The investigators interpret these fetal complications more likely as early indicators of neurodevelopmental differences rather than as causes of autism. These results emphasize prenatal origins and genetic contributions to autism and may help reduce unfounded parental guilt while guiding attention toward early detection and support.
Key facts
- Genetic and familial influence: The association between maternal health diagnoses during pregnancy and offspring autism risk is largely explained by familial confounding, including shared genes and shared environmental exposures.
- Fetal complications: The only maternal diagnoses that remained robustly associated with autism were those describing complications affecting the fetus; these are likely early signs rather than causes.
- Sibling and paternal comparisons: When mothers experienced the same diagnosis across different pregnancies, the diagnosis did not consistently predict autism risk, indicating that family-level factors explain many observed links. Similarly, comparable associations for paternal diagnoses pointed to familial explanations rather than direct prenatal effects.
Source: NYU Langone
Overview
Researchers at NYU Langone Health analyzed the medical histories of more than 1.1 million pregnancies (representing about 600,000 mothers) recorded in Denmark’s national health registries. Because Denmark links all health records to a single government identifier, the team could systematically evaluate more than 1,700 distinct ICD-10 diagnostic codes recorded for mothers before and during pregnancy. For practical analysis they focused on the 236 diagnoses that appeared in at least 0.1% of pregnancies.
Lead author Vahe Khachadourian, MD, PhD, MPH, and senior author Magdalena Janecka, PhD, used statistical models that account for a wide range of confounders, including maternal age, socioeconomic factors, disorder chronicity, and co-occurring diagnoses. Their goal was to distinguish true prenatal risk factors from associations driven by familial factors.
After adjusting for sociodemographic factors and comorbid conditions, 30 maternal diagnoses remained statistically associated with autism risk in offspring. To test whether these associations reflected causation or familial confounding, the team used two family-based approaches: discordant sibling comparisons and paternal negative-control analyses. If a maternal diagnosis is equally associated with autism when present in pregnancies of both affected and unaffected children within the same family, this points to familial rather than causal influences.
Family-based analyses showed strong evidence that the majority of observed links were attributable to familial confounding. Genetics stands out as a major familial factor: for example, genes that raise the risk of depression in a parent can also increase the risk of autism in a child, which would explain observed associations between maternal depression and offspring autism without implying a direct prenatal effect.
The study further compared maternal and paternal medical histories. If a paternal diagnosis—unlikely to affect the fetus directly—shows a similar association with autism, that pattern supports a familial explanation. Indeed, many paternal diagnoses showed associations comparable to maternal diagnoses, reinforcing the role of shared family factors.
After these family-based checks, the only maternal diagnoses that remained consistently associated with autism were obstetric diagnoses involving fetal complications. The investigators interpret these fetal diagnoses as probable early markers of atypical development rather than causal agents driving autism.
“Our analyses do not support convincing evidence that common maternal diagnoses during pregnancy cause autism,” said Magdalena Janecka. “The dominant hypothesis supported by our findings is that autism originates prenatally and is strongly influenced by genetic and familial factors. Clarifying this can ease misplaced guilt among parents and help focus clinical efforts on early identification and support for autistic children and their families.”
Autism spectrum disorder is a developmental condition typically identified in childhood and characterized by differences in social communication and repetitive or restricted behaviors. Symptoms vary widely across individuals and across the lifespan. In the United States, federal estimates indicate about one in 54 children has been identified with autism.
Funding: This work was supported by National Institutes of Health grants R01MH124817 and T32MH122394; Lundbeck Foundation grants R102-A9118 and R155-2014-1724; the Seaver Foundation; and the Eunice Kennedy Shriver National Institute of Child Health and Human Development grant HD098883.
Other co-investigators include researchers at Aarhus University (Elias Speleman Arildskov, Jakob Grove, Stefan Nygaard Hansen), Icahn School of Medicine at Mount Sinai (Paul O’Reilly, Joseph Buxbaum, Abraham Reichenberg, Sven Sandin), Kaiser Permanente Northern California (Lisa Croen), and Drexel University (Diana Schendel).
About this genetics and autism research news
Author: David March
Source: NYU Langone
Contact: David March – NYU Langone
Image credit: Neuroscience News
Original Research: Open access. “Most associations between maternal health and autism are attributable to familial confounding” by Magdalena Janecka et al., published in Nature Medicine.
Abstract (summary)
Most associations between maternal health and autism are attributable to familial confounding
This cohort included all children born in Denmark from 1998 through 2015 (n = 1,131,899) and their parents. Cox proportional hazards models estimated the risk of autism associated with each maternal prenatal ICD-10 diagnosis, accounting for chronicity, comorbidity, familial correlations, and sociodemographic factors. Family-based methods—discordant sibling comparisons and paternal negative controls—were used to evaluate familial confounding. By the end of follow-up, 18,374 (1.6%) individuals had a diagnosis of autism.
Across 236 maternal diagnoses tested (prevalence ≥ 0.1%), 30 were significantly associated with autism after accounting for confounders and multiple testing. These included obstetric, cardiometabolic, and psychiatric disorders (for example, diabetes in pregnancy and depression). Family-based analyses provided strong evidence that most observed associations were attributable to familial confounding. Overall, the findings indicate widespread associations between maternal health in pregnancy and offspring autism, but they underscore that these associations are largely explained by familial factors rather than direct causal effects of maternal diagnoses.