30502
Interaction between Maternal Immune Activation and Antibiotic Use on Subsequent Risk of Autism Spectrum Disorder
Maternal immune activation (MIA) is emerging as a consistent risk factor for ASD. However, the specific sources of MIA or the mechanisms underlying MIA-ASD association are unclear. Animal literature suggests that the maternal gut microbiota interact with her immune system to influence fetal neurodevelopment and manifest autism-like symptoms in the offspring. Antibiotic use modifies the maternal gut microbiota, but its role in modifying the MIA-ASD association is unknown.
Objectives:
We assessed whether MIA and antibiotic use during pregnancy interact to influence the risk of ASD in the offspring in a prospective birth cohort.
Methods:
Participants include 3,123 mother-child pairs from the Boston Birth Cohort, a prospective birth cohort enriched for preterm birth recruited and followed at the Boston Medical Center. MIA during pregnancy was assessed using electronic health records (EHR) reports of infection and a postpartum maternal questionnaire. Antibiotic use was assessed using EHR outpatient/inpatient prescription drug information during pregnancy, chart extraction during labor and delivery (intrapartum), and a postpartum maternal questionnaire. ASD diagnosis was based on ICD-9 diagnostic codes. Typically developing children (TD) did not have ASD, ADHD, or other intellectual/developmental disabilities. We used logistic regression to estimate potential interaction between MIA and antibiotic use on odds of ASD. Models were adjusted for maternal age, maternal education, maternal race, marital status, child sex, child’s year of birth, delivery type, preterm birth, low birth weight, and smoking during pregnancy.
Results:
At delivery, mothers were 28.5 years old on average. Most were black/African-American (58%) and had high school education level of higher (72%). Five percent of the offspring had ASD, of which 74% were males. In our multivariable logistic regression models, we found a significant interaction between flu in the second trimester and EHR-abstracted antibiotics during pregnancy on ASD risk (p for interaction=0.03). When stratified by antibiotic use, flu in second trimester was only a risk factor for ASD among mothers who did not receive an antibiotic during pregnancy (OR=6.8, p=0.09 vs. OR=1.2, p=0.74). We also found a potential interaction between fever in the second trimester and intrapartum antibiotic use (p for interaction=0.07). When stratified by intrapartum antibiotic use, fever remained a risk factor only among the mothers who did not receive an intrapartum antibiotic (OR=3.7, p=0.01 vs. OR=0.66, p=0.58). We did not find evidence of interaction in the first or third trimester between flu/fever and antibiotic use.
Conclusions:
In this enriched risk cohort study, we found evidence of a potential interaction between second-trimester flu or fever and antibiotics during pregnancy on ASD risk. Offspring with maternal flu/fever who did not receive antibiotics were at highest risk. We did not find evidence of interactions in the first or third trimester, or overall pregnancy. This study builds on existing animal literature demonstrating an interaction between MIA and antibiotic use and ASD-like symptoms in the offspring. More work is needed to replicate this finding.