The use of prescription opioid analgesic (POA) medications during pregnancy is increasing.1 Studies have linked maternal POAs during pregnancy to adverse birth outcomes in offspring (e.g., preterm birth).2 However, research on later childhood outcomes is lacking. Because pregnant women are typically excluded from randomized clinical trials of medications, researchers must use observational methods to study prenatal POA exposure. Research suggests that maternal POA use during pregnancy is correlated with environmental risk factors3 and genetic liability for neurodevelopmental problems.4 Therefore, to rigorously evaluate associations between prenatal POA exposure and two neurodevelopmental problems--autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD)--we used records of POA prescriptions filled by pregnant women and observational methods that help account for genetic and environmental factors that differ between exposed and unexposed offspring.
We studied 840,659 singleton offspring born 2006 to 2013. First, we assessed associations between pregnancy, maternal, paternal, and other familial and socioeconomic factors and POA exposure. Second, we estimated associations between POA exposure and the neurodevelopmental problems while controlling for numerous measured characteristics associated with POA exposure. Third, we compared risk of the outcomes among differentially exposed siblings while adjusting for measured characteristics that may vary among siblings. By design, sibling comparisons account for all unmeasured genetic and environmental factors that make sibling similar, including influences from maternal conditions that are stable across pregnancies. We used Cox proportional hazard regression to account for right censoring of the outcomes. For sibling-comparisons, we fit fixed-effect models using stratified Cox regression.
Numerous risk factors were associated with POA exposure (e.g., maternal smoking during pregnancy odds ratio[OR]=2.03, 95% confidence interval[CI]:1.96-2.10, prenatal exposure to other psychoactive medications OR=2.80, 95% CI:2.68-2.92). After adjustment for measured covariates, POA exposure was associated with risk of ASD (hazard ratio[HR]=1.49, 95% CI:1.28-1.74) and ADHD (HR=1.99, 95% CI:1.71-2.31). The associations remained elevated in sibling comparison models, though CIs were wider (ASD HR=1.45, 95% CI:0.71-2.95; ADHD HR=2.34, 95% CI:0.98-5.56).
In sum, our results showed that (a) several risk factors were associated with POA use during pregnancy, indicating that exposed and unexposed offspring differ on several important background characteristics, (b) the associations between POA exposure and ASD and ADHD were independent of several measured risk factors, and (c) these associations remained elevated in siblings comparisons models, indicating that associations may be independent of unmeasured genetic and environmental factors that make siblings similar. However, given that CIs in sibling comparisons were wide, future studies using larger samples will need to replicate these results. Additionally, given that all observational methods have strengths and weaknesses and control for different sets of confounding factors, future research will also need to use other observational methods to further test the role of confounding factors, such as paternal POA use during the pregnancy period as a negative control for the exposure and maternal use of other pain medications during pregnancy as active comparators.
1Desai, et. al. (2014). Obstetrics and Gynecology. 2Yazdy, et. al. (2015). Journal of Pediatric Genetics. 3Smith, et. al. (2015). Maternal and Child Health Journal. 4Kato, et. al. (2008). Psychological Medicine.
Developmental Disorders (e.g. ADHD)
