Safety of Medicines in Pregnancy

We have been awarded a NIHR programme grant on Safety of Medicines in Pregnancy which includes mixed methods approaches with analysis on QResearch as well qualitative analyses. This application is for the QResearch component and separate REC approval will be sought for the qualitative analyses. At the request of the Scientific Committee, information in the revised application only covers the analyses to be done using QResearch analyses rather than the other inter-related work packages.

1.1 What is the problem?

Medication use in pregnancy is challenging as benefits of treatment need to be considered in the context of the safety of the mother and baby in addition to the risks of stopping medication if taken before pregnancy. However, safety data are almost completely lacking for most new, and many old, medications. Pregnant women are rarely included in clinical trials for new medications, post-marketing surveillance is rarely comprehensive, and robust data are not routinely collected on outcomes after medication use in pregnancy, meaning that the safety or potential risks of harmful effects to the foetus of many medications in humans are not well understood. Currently it is only possible to learn about safety or adverse maternal and fetal outcomes associated with medication use once a drug is widely available and widely used by pregnant women through notification schemes such as the Medicines and Healthcare Products Regulatory Agency (MHRA) ‘yellow card’ system which is known to be incomplete, voluntary registries which are also subject to under-reporting, very small post-marketing follow-up studies and limited analysis of routinely collected electronic health datasets. Each of these sources have major drawbacks and inherent biases which mean the data are impossible for women and clinicians to interpret.

There is currently no established mechanism to enable systematic analysis to assess clinical benefits and risks of medications used in pregnancy using large datasets of health data to minimise the biases and limitations of current approaches. Such a system could provide the key information needed, not only for policy makers and medicine regulators, but also for pregnant women and their doctors to inform decisions on whether to start, stop, continue or change their medication during pregnancy.

1.2 Why is this research important?

Many medications can be safely used in pregnancy, but we have insufficient data on which ones are safe until several years after the same medications have been used and found to be safe and effective in the non-pregnant population. Pregnant women are thus unable to benefit from their use until long after the rest of the population and may come to harm as a result. The recent introduction of COVID-19 immunisation provides such an example – the absence of pregnancy safety data led to widespread hesitancy amongst women and clinicians and preventable deaths from COVID-19 amongst unvaccinated pregnant women.

Conversely, some medications used during pregnancy may be associated with increased risks of adverse outcomes. In 2020, the 'First do no harm’ report was published by Baroness Cumberlege, who recommended that teratogenic medicines in pregnancy (those with the ability to cause defects in a developing fetus) should be avoided wherever possible and, where prescription is unavoidable, there is a duty to ensure women are fully informed of both benefits and risks. The report followed identification of harms from in utero valproate exposure over many years which had not been formally quantified, and their extent not recognised, due to an absence of robust data systems. To determine whether a medication is safe to use in pregnancy, robust evidence of benefits and risks of the medication is needed, and a way to easily communicate these to pregnant women. However, the absence of well-conducted research studies evaluating medication use in pregnancy means that there remain significant gaps in evidence about the safety of many medications. This lack of safety evidence can lead to women inappropriately stopping medications during pregnancy or not starting treatments that may be beneficial due to safety concerns. Maternal deaths have resulted from medication withdrawal.

There is therefore a pressing need for a mechanism to conduct robust research studies into the benefits and possible harms of medication use in pregnancy to better inform clinical decisions as well as to implement the recommendations of the Cumberlege report.

1.3 Review of existing evidence.

Research into the safety and the clinical risk/benefit profile of medications used in pregnancy has been limited. Recruitment of pregnant women into randomised controlled trials has only been possible in the last 25 years and it has been limited due to uncertainty around the risk of harm to their unborn child. Thus, knowledge gained of medication efficacy and safety in pregnancy has largely been through post marketing surveillance studies. In the UK, post marketing surveillance has relied on information available through a myriad of datasets, including primary care and secondary care databases, as well as national registries such as the UK and Ireland Epilepsy and Pregnancy Register and national surveillance services such as the UK Teratology Information Service and MHRA yellow card system. As these datasets are not currently linked, it has not been possible to comprehensively follow-up women throughout their pregnancy and beyond nor has it been feasible to extensively assess their pregnancy complications and long-term outcomes in their children in the UK.

Medication use during pregnancy is common, with a recent survey in the UK identifying 76.4% of women having had at least 1 over-the-counter or prescribed medication during their pregnancy. Medications are taken for several reasons across pregnancy, ranging from treating an acute illness, such as nausea to managing long-term conditions, such as depression, diabetes and epilepsy.

Depression is a chronic condition treated in primary care. For many types of widely-used antidepressant medication, there are little data on safety of use in pregnancy. Some medications, including selective serotonin reuptake inhibitors and benzodiazepines, used to treat mood disorders, have been associated with poor maternal pregnancy outcomes, such as a higher risk of postpartum haemorrhage, poor perinatal outcomes, such as stillbirth and neonatal intensive care admissions, and potential increased risk of mood disorders in childhood. However, confounding due to having the health condition itself rather than the medication has led to difficulties identifying causation. There are few studies comparing risks of adverse outcomes in women taking antidepressant medication for depression during pregnancy to risks in women with untreated depression, which is the information needed to weigh the benefits and risks of treatment.

Nausea and vomiting are common in pregnancy. If symptoms are severe, it can result in dehydration or malnutrition. The burden of the severe forms is known to be highest in women from ethnic minorities and low socioeconomic groups. Antiemetic medications are prescribed in 49% of pregnant women with nausea and vomiting who present to their GP, often in early pregnancy. Some antiemetic medications have been assessed for safety in pregnancy and there are reported associations with cleft palate and major congenital abnormalities. However, there is some debate about whether risks of treatment outweigh the benefits, and there is a recognition that there is generally a paucity of data.

Some anti-epileptic medications, including sodium valproate, are known teratogens, but continue to be used by women of child-bearing age. A study in over 7000 pregnant women using a European database, determined that seven epilepsy medications had an increased risk of congenital abnormalities. Furthermore, although the evidence is limited, antiepileptic use in pregnancy has been associated with poorer neurodevelopmental outcomes in offspring. Overall, there remains a large amount of uncertainty in the safety of anti-epileptic medications on outcomes because existing data are from studies that are of poor quality or too small. There are no United Kingdom (UK)-based data on risks of epilepsy treatment, or of untreated epilepsy, in pregnancy.

Approximately 1-2% of pregnant women have pre-existing diabetes and a further 6-8% develop it during pregnancy. Hence, the safety of diabetes medications during pregnancy is crucial to ensure both maternal and fetal health. Insulin is considered the safest treatment for managing blood sugar in pregnant women with type 1, type 2, or gestational diabetes, as it does not cross the placenta. Metformin and glyburide, commonly used for type 2 diabetes, are sometimes used, but their long-term effects on fetal development are still not fully understood. Metformin is generally considered safe, though some concerns remain about fetal growth. there is limited large-scale trial data for newer classes like GLP-1 receptor agonists or SGLT2 inhibitors, which are generally avoided during pregnancy due to potential risks. Careful investigation of different types of medicines used to treat diabetes during pregnancy is therefore essential.

This programme of work will link prescription, secondary care and adverse event reporting databases to the QResearch primary care database, one of the largest consolidated primary care research databases in the UK with over 1500 GP practices. It will use this novel methodology for investigating post-marketing medication safety in pregnant women to provide a mechanism to close the evidence gaps for treatment of health conditions in pregnancy.

GP data - Demographics (year of birth, age, ethnic group, deprivation); co-morbidities (e.g. diabetes, hypertension); medication; family history; pregnancy measurements and risk factors; pregnancy outcomes;

Hospital Episode Statistics - Admitted Patients - date admission and discharge, type of admission, destination on discharge, clinical codes

Hospital Episode Statistics - Outpatients - date appointment, clinical codes

Hospital Episode Statistics - Critical Care - Hospital Episode Statistics - Critical Care

COVID-19 Infection (SGSS)- date and code of covid infection

COVID-19 Vaccination - COVID-19 Vaccination

Maternity Services (MSDS) - Maternity Services (MSDS)

Congenital Abnormalities - date, type of abnormality

Civil Registration (Mortality) - date of death, up to 15 causes of death, underlying cause of death

Cancer Registry - cancer diagnoses and treatment

Professor Carol Coupland (University of Nottingham)

Dr Defne Saatci (University of Oxford)

Dr Jenny Hirst (University of Oxford)

Professor Marian Knight (University of Oxford)