No relationship between labor induction and an increased risk for autism spectrum disorders (ASD) is seen when family variables are taken into account, a new study suggests. These findings run counter to previous studies, in which an association between induction and ASD was reported.

The difference lies in the use of a family comparison design involving discordant pairs of siblings or first cousins; that is, comparing one child born after induction of labor with a relative born after no induction of labor, lead author Anna Sara Oberg, PhD, and colleagues write in an article published online July 25 in JAMA Pediatrics. This allowed the authors to “control for all shared maternal factors (present across all pregnancies) that are unmeasured in registries but appear to confound the association between labor induction and neurodevelopmental disorders in the offspring.”

The findings suggest that concerns about ASD “should not factor into the clinical decision about whether to induce labor,” they write.

Dr Oberg, from the Department of Epidemiology, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, and coauthors studied all live births in Sweden between 1992 and 2005. Using several nationwide registries that include all Swedish residents, they calculated induced births, ASD diagnoses, and maternal lifestyle and socioeconomic information such as reproductive history, use of tobacco, health history, and cohabitation status. The authors followed all the children born during the study period through the end of 2013 or until they were diagnosed with ASD, died, or emigrated from Sweden.

The baseline analysis accounted for birth year, parity, and maternal age, in addition to induction of labor. The researchers then added other covariates, including stable ones such as maternal educational level, as well as those specific to each pregnancy, such as maternal smoking, multiple gestations, preeclampsia, and urogenital infection, among others. The final model included a “fixed-effect” adjustment to allow for comparison between maternal siblings or maternal first cousins, while continuing to account for the covariates that were unique to each birth.

There were 1,362,950 members of the cohort, including 22,077 who were diagnosed with ASD during follow-up. Labor induction was used in 11% of live births during the study. Of the maternal sibling pairs in the sample, 15.2% were discordant for labor induction, as were 18.2% of the maternal cousin pairs. Primiparity, older maternal age, and higher maternal body mass index all were risk factors for induction, along with pregnancy complications such as gestational diabetes, gestational hypertension, and preeclampsia. By the time the cohort members were 20 years of age, ASD had been diagnosed in 3.5% of the offspring in the induced sample and 2.5% of the offspring in the noninduced sample.

The initial analyses appeared to confirm the findings of earlier studies: Labor induction was associated with a significantly higher risk for ASD (hazard ratio [HR], 1.32; 95% confidence interval [CI], 1.27 – 1.38) in the baseline model. This did not change substantially after adjustment for stable maternal characteristics (HR, 1.31; 95% CI, 1.26 – 1.37). Additional adjustment for pregnancy-specific factors resulted in a slight reduction in association (HR, 1.19; 95% CI, 1.13 – 1.24), but the association was still significant. However, the association disappeared (HR, 0.99; 95% CI, 0.88 – 1.10) when the authors applied fixed-effects models, “comparing discordant siblings to each other to account for all the factors they share.”

These findings suggest there is some other, still-unknown factor responsible for confounding the relationship between labor induction and ASD seen in earlier studies, Dr Oberg and colleagues write. Genes that govern cellular calcium homeostasis might be one culprit. An environmental factor might be delivery at a higher-intensity medical system, where clinicians might induce labor more readily and diagnose neurodevelopmental disorders more frequently.

Several prenatal and perinatal factors have been studied as possible candidates for ASD risk, but often the result is simply more questions, Daniel L. Coury, MD, writes in an editorial accompanying the article. For example, some evidence implicates maternal use of selective serotonin reuptake inhibitors in the incidence of ASD. “Should pregnant women take these medications, or should physicians advise against?” he asks. Similarly, one study showed a higher risk for intellectual disability among children conceived through assisted reproductive technology, but a later study, which followed a different cohort of children only through 36 months of age, failed to support this finding. “Which study is correct? How should clinicians counsel families regarding the risk?”

To help parents make better decisions, Dr Coury, from the Section of Developmental-Behavioral Pediatrics, Department of Pediatrics, Nationwide Children’s Hospital/The Ohio State University, Columbus, suggests more extensive discussion of research findings in lay terms and weighing the benefits against the risks. “The suicides prevented by [selective serotonin reuptake inhibitor] medications outweigh most concerns of adverse effects. The potential that each child brings to the world outweighs any risk associated with [assisted reproductive technology]. The benefits of labor induction, when performed in accordance with clinical guidelines, include the delivery of a healthy neonate and a healthier outcome for the mother.”

Study limitations include a lack of information on the type of labor induction used and the inability to identify the factors responsible for raising the risk for ASD in this population, the authors write. Still, they conclude, “the findings of this study provide no support for a causal association between induction of labor and offspring development of ASD.”