Objectives (1) To evaluate the direct (un-mediated) and indirect (mediated) relationship between antenatal exposure to opioid agonist medication as treatment for opioid use disorder (MOUD) and the severity of neonatal opioid withdrawal syndrome (NOWS), and (2) to understand the degree to which mediating factors influence the direct relationship between MOUD exposure and NOWS severity. Methods This cross-sectional study includes data abstracted from the medical records of 1294 opioid-exposed infants (859 MOUD exposed and 435 non-MOUD exposed) born at or admitted to one of 30 US hospitals from July 1, 2016, to June 30, 2017. Regression models and mediation analyses were used to evaluate the relationship between MOUD exposure and NOWS severity (i.e., infant pharmacologic treatment and length of newborn hospital stay (LOS)) to identify potential mediators of this relationship in analyses adjusted for confounding factors. Results A direct (un-mediated) association was found between antenatal exposure to MOUD and both pharmacologic treatment for NOWS (aOR 2.34; 95%CI 1.74, 3.14) and an increase in LOS (1.73 days; 95%CI 0.49, 2.98). Delivery of adequate prenatal care and a reduction in polysubstance exposure were mediators of the relationship between MOUD and NOWS severity and as thus, were indirectly associated with a decrease in both pharmacologic treatment for NOWS and LOS. Conclusions for Practice MOUD exposure is directly associated with NOWS severity. Prenatal care and polysubstance exposure are potential mediators in this relationship. These mediating factors may be targeted to reduce the severity of NOWS while maintaining the important benefits of MOUD during pregnancy. Significance What is already known on this subject? The use of MOUD during pregnancy improves fetal outcomes, while antenatal exposure to MOUD increases the risk of NOWS. The severity of NOWS is influenced by MOUD type, co-exposures, adequacy of prenatal care, and the infant’s gestational age. What this study adds? This study identifies factors that mediate the direct influence of antenatal MOUD exposure on the severity of NOWS and quantifies the degree to which this mediation influences outcomes in a large and geographically diverse population. Thus, providing clinicians with potential targets to improve care for this vulnerable population.