This tender is composed of two objectives:
1. Objective 1: To describe utilisation of antiepileptics in individuals of childbearing age, pregnant, and male individuals. This objective will be described in WorkPackage (WP) 1. There is a specific protocol for WP1. The development of WP1 protocol is one of the deliverables of the project.
2. Objective 2: To assess the feasibility of estimating the risk of adverse pregnancy, neonatal and child outcomes following in utero and paternal exposures to antiepileptic drugs. This objective will be assessed in WP2. Initially, only databases with developed father-child linkage developed were expected to participate. In the end, EMA required all databases involved in WP1 to participate also in WP2. WP2 does not require a data extraction, as it will answer the question whether data sources are fit for purpose for studies assessing adverse pregnancy, neonatal and child outcomes following either ASM exposure through the mother, or periconceptional ASM exposure through the father. There is a specific protocol for WP2. The development of WP2 protocol is one of the deliverables of the project.
Prototol 1 – WP1: THE UTILISATION OF ANTISEIZURE MEDICATIONS IN MEN, WOMEN OF CHILDBEARING AGE, AND PREGNANT WOMEN: A MULTI-DATABASE STUDY FROM 7 EUROPEAN COUNTRIES.
Background: Currently there is substantial evidence on teratogenicity of several antiseizure medications (ASMs) and a potential increased risk of neurodevelopmental disorders in young children exposed in utero to ASMs. Recently, few evidence has emerged expanding this risk also when there was a paternal exposure before conception. On that line, EMA asked for an updated evaluation on trends in use of various ASMs and related drugs in pregnant women, in other women of childbearing potential, and in men.
Objectives: The main objective of this study is to describe the utilisation of ASMs and related drugs (i.e., antiepileptics (ATC codes N03A), gabapentinoids (N02BF), and all benzodiazepines with antiepileptic properties) in pregnant women, other women of childbearing potential (12-55 years of age), and men (12 years and older). It has the following sub-objectives: 1.1) To estimate the annual incidence and prevalence rate of ASM use in women of childbearing potential and in men; 1.2) To describe treatment duration, discontinuation, and treatment switches to other ASMs or alternative medications and polytherapy in women of childbearing potential and men; 1.3) To estimate incidence and prevalence of specific ASM use in up to one year prior to and during the pregnancy period in all pregnant women; 1.4) To estimate and compare discontinuation, treatment switches to other ASMs or alternative medications and polytherapy among pregnant women in the 12-10, 9-7, 6-4, and 3 months prior to pregnancy or during first, second and third pregnancy trimesters; 1.5) To estimate dose changes of ASMs in women prior to and during pregnancy.
Study Design: Retrospective population-based cohort study.
Exposure and outcomes: The main exposures of interest are ASMs (N03A), including gabapentinoids (N02BF) and benzodiazepines with antiepileptic properties. The drug utilisation outcome measures will be the incidence and prevalence of ASM use among pregnant women, other women of childbearing potential, and men, treatment duration of ASM, discontinuation of ASMs, initiation of ASM use during pregnancy, switching to another ASM or an alternative medication, dose changes of ASMs before and during pregnancy, and polytherapy of ASMs.
Data Analysis: According to sub-objectives stated above, statistical estimates will be produced, such as descriptives (counts, percentages), distributions (mean, percentiles), rates (incidence, prevalence), or other relevant estimates. Various visualisations (e.g., flow diagrams, line and bar charts, and Sankey diagrams) will be used to depict the findings. This will be conducted using a common R script.
Limitations: Our operational definitions of treatment episodes with a grace period of 30 days, temporary break (between 30-120 days of no new prescription/dispensing), ASM discontinuation (no new record after 90 days), and switching from an ASM to another ASM or an alternative medicine might not reflect the exact clinical scenarios in case of all ASMs, especially when using data from diverse EHDs. One especial case would be misclassification of prevalent users as incident users due to considering not long enough look-back periods. Due to this, we will not estimate incident use of ASMs among women and men of childbearing potential and initiation of ASM during pregnancy in data sources EFEMERIS and THL that we only have 2.5 or 3-months look-back data for medications. But the overall approach is based on prior drug utilisation studies on the same topic,8,28 and same definitions for all ASMs across all centres, which considering the limitations and data quality will assist in comparability of findings. Another limitation could be operational definitions for high, medium, and low daily doses of ASMs based on DDD data, which might not always reflect the actual clinical situation. But considering the diversity of data sources and healthcare systems involved, this seems the only feasible option in such a multi-database study. Also, capturing the right doses would be difficult when prescriptions are not renewed but patients were using the old prescription with new instructions from clinician, which is not visible in all our included data sources. This will be taken into account when interpreting findings from the dose analyses.
Protocol 2 – WP2: FEASIBILITY OF ESTIMATING THE RISK OF ADVERSE PREGNANCY, NEONATAL AND CHILD OUTCOMES FOLLOWING EITHER IN UTERO ASM EXPOSURE THROUGH THE MOTHER, OR PERI-CONCEPTIONAL ASM EXPOSURE THROUGH THE FATHER
Rationale and background: Evidence shows that certain ASMs such as valproate, pose teratogenic and neurodevelopmental risks during pregnancy. Conflicting findings also suggest potential neurodevelopmental impacts from paternal ASM exposure pre-conception. Thus, it is critical to assess the feasibility of causal studies on ASM exposure, especially given the complexity of father-child data linkage.
Research question and objectives: This study aims to understand whether data sources can be used to study the effects of maternal and paternal exposure to ASM, on pregnancy, neonatal and child outcomes. For this main objective, the following sub-objectives will be addressed:
a. To estimate the availability of relevant information/characteristics for pregnant women, using 15 different parameters that will inform the assessment of fitness for purpose.
b. To estimate availability of relevant information/characteristics for men and linkage with pregnancies, using 9 different parameters that will inform the assessment of fitness for purpose.
c. To estimate availability of relevant information/characteristics for neonates/children, using 17 parameters and comparing them between those that can and cannot be linked to mother and/or father where possible.
d. To assess fitness for purpose to different types of studies of pre-conceptional/prenatal exposure to antiepileptics and the development of adverse pregnancy and child outcomes.
Study design: This is a retrospective cohort study to assess the suitability of the available databases for ASM exposure studies, aligning with the study’s fit for purpose evaluation.
Population: The source population comprises all persons of childbearing age who are registered with the data sources that participate in this study. From the source population, we will select study cohorts for each of the different sub-objectives.
Variables: The main exposures of interest are ASMs (N03A) and gabapentinoids (with ATC code N02BF) and all benzodiazepines with antiepileptic properties. Outcome parameters will be considered per subobjective. Clinical outcomes of interest will include pregnancy outcomes (i.e., spontaneous pregnancy loss, stillbirth, preterm birth) neonatal/child outcomes (e.g., small gestational age, congenital anomalies, adverse neurodevelopment).
WP2 does not require a data extraction, as it is a study to assess the fit for purpose of the databases for assessing adverse pregnancy, neonatal and child outcomes following either ASM exposure through the mother or periconceptional ASM exposure through the father.
Data analysis: For objective a-c we will estimate the 41 feasibility parameters using descriptive analyses and visualizations. For objective d, we will use the parameters from objectives a-c plus metadata on the data sources and assess the fitness-for-purpose of the data instance by using, implementing, and adapting the framework from Gatto and colleagues.
Limitations: Methods for descriptive analysis have been employed in previous studies undertaken across the ConcePTION project, EU PE&PV and VAC4EU networks (Durán et al., 2023a; Hurley et al., 2023; Abtahi et al., 2023; Durán et al., 2023b). This protocol aims to leverage existing tools available tools and algorithms, developed by the aforementioned enterprises, including the ConcePTION CDM, the pregnancy algorithm, the INSIGHT quality checks, code lists and algorithms, as well as component analyses and the analytical pipeline and functions. Differences between data sources will be described transparently. The aim is to describe and appraise mother-child and father-child linkage, both deterministic and probabilistic, that has been established thus far across each data source. We also aim to investigate novel improvement in linkage using additional parameters available in EHR data. It is possible that improvements in probabilistic linkage will not be possible due to characteristics of available data fields, or because linkage will not improve materially with inclusion of these additional data fields. Exploration of parental linkage will be described transparently, regardless of whether improvement in linkage is possible. Due to limited budget and the scope of work, we will not be able to validate any of the linkages nor outcomes.