Psychotropic medicines work by adjusting and balancing the levels of important brain chemicals and are prescribed to treat a variety of mental health conditions. However, the prescribing of psychotropic medicines to pregnant women is avoided wherever possible because they can affect the development of, and cause harm to, the baby in the womb. Despite this, sometimes psychotropic medication cannot be avoided because it is required to treat short or long term mental health conditions during pregnancy. Harmful effects on the baby, which are immediately obvious at birth, have been reported for these medicines. Less is known however about the type of longer-term harm they may cause the baby in the womb, which may only be obvious in later years.
Data from the Clinical Practice Research Datalink (CPRD) include information on births and other maternity events, subsequent pregnancy and child outcomes, prescriptions received, diagnoses recorded in GP practices, hospitalisations, and deaths. Accessing CPRD data will enable our team to analyse these records linked together for the same mother and subsequent records linked together for her children to address important gaps in understanding and inform clinical practice. Our team aim to investigate the association between psychotropic medications taken during pregnancy and:
1. pregnancy outcomes
2. subsequent child health outcomes
We will investigate pregnancy and child outcomes for women prescribed psychotropic medications during pregnancy (medications within BNF Chapter 4 including those prescribed for anxiety [BNF 4.1], psychosis [BNF 4.2], depression and mood disorder [BNF 4.3], ADHD [BNF 4.4], and epilepsy [BNF 4.8]) compared to a) women with the underlying conditions of interest who did not receive corresponding psychotropic medication and b) women who did not have these underlying conditions and did not receive psychotropic medication. We will ascertain women with pregnancies ending between 1st January 1991 and 31st December 2020 and link mother's Clinical Practice Research Datalink (CPRD) primary care and prescribing data to their children's CPRD primary care and prescribing, Hospital Episode Statistics (HES) inpatient admission, and Office for National Statistics (ONS) death data.
Outcomes will be analysed using:
Binary logistic regression – spontaneous abortion (y/n), fetal death (y/n), stillbirth (y/n), infant death (y/n), congenital anomaly (y/n), specific types of congenital anomaly (y/n), low birthweight (y/n), preterm birth (y/n), asthma (y/n), diabetes (y/n), autism (y/n), ADHD (y/n), depression (y/n), other neurodevelopmental conditions (y/n).
Ordinal logistic regression – birthweight categories, gestational age categories, Apgar score categories.
Multinomial logistic regression – mode of delivery.
Poisson regression / negative binomial regression – number of hospital admissions, length of hospital stays.
Cox proportional hazards regression / Poisson piecewise regression – hospitalisation, mortality.
Primary care data will enable identification of untreated disease to differentiate effects of medication from effects of the underlying condition. Where powered, we will do sub-group analyses of medicated women: continued medication during pregnancy; stopped medication during pregnancy; and started medication during pregnancy. We will investigate any critical time of exposure (1st, 2nd or 3rd trimester) and relationships with dose or duration of medication.
We will adjust for confounders detailed in the 'exposures, outcomes and covariates' section.
Pregnancy outcomes of interest include spontaneous abortion, fetal death and stillbirth.
For live births, birth outcomes include premature birth, low birthweight, mode of delivery, Apgar score, congenital anomalies and infant death, whilst child outcomes include increased hospitalisation and mortality, chronic conditions such as asthma and diabetes, and neurodevelopmental conditions such as autism and ADHD.
Michael Fleming - Chief Investigator - University of Glasgow
Michael Fleming - Corresponding Applicant - University of Glasgow
Claire Hastie - Collaborator - University of Glasgow
Daniel Mackay - Collaborator - University of Glasgow
David Blane - Collaborator - University of Glasgow
James McLay - Collaborator - University of Aberdeen
Jill Pell - Collaborator - University of Glasgow
CPRD Mother-Baby Link;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Pregnancy Register