Abstract Background Congenital heart defects (CHDs) represent the most common congenital anomalies, and a significant proportion is suspected to be caused by foetal drug exposure. Maternal medication use during pregnancy is prevalent, but most research focuses on single agents, often ignoring the potential influence of polypharmacy and timing during foetal development. Novel analytical strategies that integrate medication type, exposure duration, and critical windows of foetal development may offer deeper insights into how specific exposure patterns affect CHD risk. Purpose We applied a novel data-driven machine learning algorithm to characterize real-world maternal medication use patterns during the periconceptional period and evaluated their association with major CHDs in the offspring. Methods This register-based nationwide cohort study included all live births from 1997 to 2023. Prescription data captured all maternal prescriptions filled from 12 weeks before conception through gestational week 13. Hierarchical cluster analysis (using TAME package, R software) identified medication use patterns based on the number, pharmacological class (Anatomical Therapeutic Chemical codes), and timing of filled prescriptions. Registered offspring CHDs were identified using International Classification of Diseases, 10th Revision, codes and classified as major if they required surgical or catheter-based intervention within the first year of life or were the primary cause of death. Logistic regression models estimated adjusted odds ratios (aORs) for major CHDs, comparing users of each cluster of medications with a nonexposed reference group and adjusting for maternal age, employment status, education, urbanization, parity, multiple pregnancy, gestational age, birth weight, and further adjusted for smoking and body mass index in sensitivity analyses. Results Among 1,579,543 liveborn children, 20,378 (1.29%) were diagnosed with a major CHD. Eight distinct maternal medication use clusters were identified. Three medication use patterns were associated with a significantly increased risk of major CHDs compared with nonexposed pregnancies: 1. intensive polypharmacy (≥4 concurrent medications), primarily combining fertility treatments and anti-infective medications (odds ratio OR 1.40, 95% confidence interval CI 1.14–1.72); 2. monotherapy with anti-infective medications (OR 1.06, 95% CI 1.02–1.10); and 3. monotherapy with antacids, antiemetics, antihistamines, or analgesics (OR 1.06, 95% CI 1.01–1.11) (Figure 1). Conclusions Intensive maternal polypharmacy with fertility treatments and anti-infectives was mostly strongly associated with the risk of major CHDs, increasing the risk by 40%. These novel findings underscore the importance of refined exposure classifications. Also, the findings call for further assessments of the underlying mechanisms to clarify whether changes in clinical practice may protect exposed children from CHD.
Ninh et al. (Sat,) studied this question.