Maternal medication use patterns before and during pregnancy and cardiac outcomes in newborns
G Ninh, A D Laksafoss, R O Voegg, A S Sillesen, A A Raja, U L Beauchamp, A F Mariager, C Torp-Pedersen, A Pinborg, M P Andersen, J T Andersen, H Bundgaard, H A Boyd, K K IversenAbstract
Background
Maternal prescription medication use during pregnancy is common. Most studies assess use of single medications using binary exposure definitions and focus on major congenital heart defects (CHDs). Thus, the impact of real-world, time-varying maternal medication use patterns on newborn cardiac structure and function remains uncertain.
Purpose
To identify maternal medication use patterns from 12 weeks before conception through delivery and evaluate their associations with newborn transthoracic echocardiographic (TTE) parameters and the most frequent minor CHDs.
Methods
In a population-based cohort of 24,245 term newborns who underwent TTE within 60 days of birth, filled maternal prescriptions were ascertained from a national prescription registry. Unsupervised hierarchical clustering analyses identified exposure patterns based on medication class (Anatomical Therapeutic Classification codes), timing of use, and polypharmacy. Associations with 13 echocardiographic parameters were assessed using linear regression (mean differences) and quantile regression (10th, 25th, 75th, 90th percentiles). Logistic regression evaluated minor CHDs detected on neonatal TTE, including ASD, VSD, and BAV. Models were adjusted for maternal and neonatal covariates with correction for multiple testing.
Results
Overall, 60.7% of newborns were exposed to at least one prescription medication before or during pregnancy. Eight distinct medication-use patterns were identified. Across patterns, adjusted differences in cardiac structure and function were small and showed no consistent direction, and no pattern was associated with ASD, VSD, or BAV. One pattern characterized by sustained use of psychiatric medications, including selective serotonin reuptake inhibitors, together with analgesics was associated with small differences at the lower end of the distribution, including larger left ventricular internal diameter in diastole (25th percentile: 0.23 mm, 95% CI 0.04,0.42), larger left ventricular end-diastolic volume (25th percentile: 0.36 mL, 95% CI 0.06,0.67), and shorter mitral deceleration time (10th percentile: -5.08 ms, 95% CI -7.29, -2.88).
Conclusion
Common maternal medication use patterns before and during pregnancy were not associated with clinically meaningful differences in neonatal cardiac structure or function or increased risk of minor CHDs. Small differences observed at the lower end of the distribution in one medication-use pattern warrant further longitudinal evaluation.