New study challenges link between mother's health during pregnancy and autism risk

While many studies have reported a link between a mother's health condition during pregnancy and her child's risk of autism, a new study shows that nearly all of these "associations" can otherwise be explained by factors such as genetics, exposure to pollution, and access to healthcare.

Led by researchers at NYU Langone Health, the study revealed that, of the few conditions truly associated with autism, all were actually complications with the fetus - leading the authors to believe that those symptoms were early signs of autism in the child and not the cause of it. 

"Our study shows that there is no convincing evidence that any of these other diagnoses in the mother can cause autism," said study senior author Magdalena Janecka, PhD, an associate professor in the Department of Child and Adolescent Psychiatry and in the Department of Population Health, at NYU Grossman School of Medicine. 

Publishing in the journal Nature Medicine online Jan. 31, the new study included an analysis of 
the medical histories of more than 1.1 million pregnancies (among 600,000 mothers) from a national registry in Denmark. Unlike medical records in the United States, which are often scattered among many different medical providers an individual sees during their lifetime, in Denmark all of an individual's health records are consolidated under a single government-issued number, which enabled researchers to check each woman for more than 1,700 distinct diagnoses as defined by international standards, known as ICD-10 codes. From these, researchers focused their analysis on those diagnosed in at least 0.1% of pregnancies (236 diagnoses).

We believe our study is the first to comprehensively examine the entire medical history of the mother and explore a wide range of possible associations, controlling for multiple concurrent conditions and confounding factors."

Vahe Khachadourian, MD, PhD, MPH, study lead author, research assistant professor in the Department of Child and Adolescent Psychiatry at NYU Grossman School of Medicine

Denmark has strict safeguards in place to prevent misuse of the registry data, says Janecka, since it contains personal information. But because of the individual-specific information, the researchers were able to cross-check every diagnosis a woman had had with her children's risk of autism.

For the study, the researchers corrected for factors that could confound, or offer an alternative explanation for, the link between the diagnosis a woman received and a child's autism diagnosis. These factors include sociodemographic status and the mother's age during pregnancy, since children of older mothers are more likely to be diagnosed with autism, and their mothers are also more likely to receive certain diagnoses, such as hypertension, than their younger counterparts. 

After accounting for these confounding factors, as well as for concurrent diagnoses, 30 were still statistically associated with autism in the child. To determine if these happened to occur alongside rather than cause autism, the researchers then included the siblings of autistic children in the analysis. If a mother was diagnosed with the same condition during pregnancies of children with and without autism, then it would suggest that factors other than her diagnosis were influencing the link with autism. This step disentangled the conditions that could be attributable to familial factors, such as genetics and environmental exposure to pollution, from those that may be causing autism. 

Genetics is a strong familial confounder (plausible explanation) for autism, the researchers say. Certain genes that increase the risk of someone having depression are also more closely tied to them having autism. If a woman suffers a bout of depression during pregnancy and her child is autistic, it is much more likely that mother and child share genes that cause both conditions, rather than that the chemical effects of depression somehow affected the fetus to cause autism during development. 

Researchers also analyzed fathers' medical histories. Any association between a paternal diagnosis and autism would most likely be caused by familial factors, since the father's direct effects on a fetus postconception are likely very limited. In fact, the researchers observed that a lot of paternal diagnoses are just as related to child autism as the maternal diagnoses.

After accounting for the familial factors, the only maternal diagnosis that was still strongly statistically associated with autism was pregnancy complications related to the fetus. 

"Our interpretation is that these fetal diagnoses likely do not cause autism, but are instead early signs of it," said Janecka. "The predominant hypothesis is that autism really starts prenatally. Even before a child receives a diagnosis for autism, developmental changes have been happening the entire time. 

"Many mothers of children with autism feel guilty about it," said Janecka, "thinking that they did something wrong during pregnancy, and it is heartbreaking. I think showing that these things are not going to cause autism is important and may lead to more effective ways to support autistic children and their families."

Autism is recognized as a developmental disorder that often appears in childhood and is marked by a range of difficulties with social interactions and repetitive behaviors. Symptoms vary widely into adulthood but can include reduced eye contact, reluctance to engage in playtime activities, repeating gestures or sounds, and an indifference to temperature extremes. According to federal estimates, one in every 54 children in the United States is affected by autism.

Funding support for the study was provided 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 study co-investigators are Elias Speleman Arildskov, Jakob Grove, and Stefan Nygaard Hansen at Aarhus University in Denmark; Paul O'Reilly, Joseph Buxbaum, Abraham Reichenberg, and Sven Sandin at Icahn School of Medicine at Mount Sinai in New York City; Lisa Croen at Kaiser Permanente Northern California in Oakland; and Diana Schendel at Drexel University in Philadelphia.