Methodological challenges in pregnancy pharmacoepidemiology : the case of antiseizure medication and offspring neurodevelopment

This thesis is driven by the overarching aim of elucidating methodological challenges inherent in pharmacoepidemiology during pregnancy, with a particular focus on antiseizure medications. In Study I, a substantial surge in antiseizure medication use in the United Kingdom is observed, notably linked to increases in psychiatric indications. This shift in the medication landscape raises questions about the predominant contributors to this rise and underscores the necessity of understanding evolving patterns in drug utilization during pregnancy. In Study II, a large-scale examination of associations between specific antiseizure medications and neurodevelopmental conditions across Sweden and the United Kingdom emphasizes the importance of considering drug classes and shared confounders, providing valuable insights into the possible causal effect of these drugs.

Study III explores the intricate interplay between epilepsy and psychiatric conditions, unveiling a heightened risk of neurodevelopmental conditions in individuals diagnosed with epilepsy. These findings shed light on the complex within-individual links between these conditions, potentially explaining the observed higher likelihood of neurodevelopmental diagnoses in children of women using antiseizure medications in pregnancy. In Study IV, a critical evaluation of drug safety studies warns against indication-based sampling, advocating for comprehensive regression adjustments to mitigate biases. Finally, Study V introduces the marginalized between-within model, a novel approach to derive absolute measures of occurrence in sibling analysis, enhancing the interpretability of findings.

This thesis, collectively, calls for a concerted effort to improve the methodology of pharmacoepidemiology during pregnancy, fostering a more nuanced understanding of medication risks, and ultimately enhancing maternal and fetal health outcomes.

List of scientific papers

I. Madley-Dowd P*, Rast J*, Ahlqvist VH, Zhong C, Martin FZ, Davies NM, Lyall K, Newschaffer C, Tomson T, Magnusson C, Rai D, Lee BK, Forbes H. Trends and patterns of antiseizure medication prescribing during pregnancy between 1995 and 2018 in the United Kingdom: A cohort study. BJOG. 2024;131(1):15-25. *Equal contributions.
https://doi.org/10.1111/1471-0528.17573

II. Madley-Dowd P*, Ahlqvist VH*, Forbes H, Rast J, Zhong C, Martin FZ, Barry CJS Barry, Berglind D, Lundberg M, Lyall K, Newschaffer C, Tomson T, Davies NM, Magnusson C, Rai D, Lee BK. Antiseizure Medication Use During Pregnancy and Neurodevelopmental Outcomes in Offspring: A Study of Electronic Health Records from the UK and Sweden. *Equal contributions. [Manuscript]

III. Ahlqvist VH, Dardani C, Madley-Dowd P, Forbes H, Rast J, Zhong C, Gardner R, Dalman C, Lyall K, Newschaffer C, Tomson T, Lundberg M, Berglind D, Davies N, Lee BK, Magnusson C, Rai D. Psychiatric comorbidities in epilepsy: population co-occurrence, genetic correlations and causal effects. [Accepted]
https://doi.org/10.1136/gpsych-2023-101201

IV. Ahlqvist VH, Madley-Dowd P, Ly A, Rast J, Lundberg M, Jonsson-Bachmann E, Berglind D, Rai D, Magnusson C, Lee BK. Bias amplification of unobserved confounding in pharmacoepidemiological studies using indication-based sampling. Pharmacoepidemiol Drug Saf. 2023;32(8):886-897.
https://doi.org/10.1002/pds.5614

V. Ahlqvist VH, Sjöqvist H, Sjölander A, Berglind D, Lee BK, Madley-Dowd P. Moving beyond risk ratios in sibling analysis: estimating clinically useful measures from family-based analysis. [Manuscript]