Abstract
Introduction
mostly in the first few weeks. Although these medications are not advised against during pregnancy for those who may need them, it is still unknown whether they causally increase the risk of adverse pregnancy outcomes, such as miscarriage or stillbirth. Whether these observed associations are causal, or an artefact of bias is yet to be confirmed. We aimed to use the Clinical Practice Research Datalink (CPRD), a UK-wide repository of primary care data, and a range of comparisons to investigate antidepressant use during trimester one and pregnancy loss.
Methods
We included pregnancies that occurred between 1996 and 2018 and among women who had been diagnosed with an indication for antidepressants (depression, anxiety or other (including migraine prophylaxis and stress incontinence). Exclusions were made based on outcome and follow-up time. Antidepressant use during trimester one was defined as having at least one prescription overlapping with or starting between the start date of pregnancy and the start of trimester two. Exposed pregnancies were also split into prevalent users (those using antidepressants before pregnancy) and incident users (those not using antidepressants before pregnancy) for a secondary analysis. Miscarriage and stillbirth were defined in the Pregnancy Register, supplemented with Hospital Episode Statistics (HES) where possible. Hazard ratios and 95% confidence intervals were generated using multivariable Cox proportional hazards models. Additional analyses are ongoing, including a propensity score analysis and exposure discordant pregnancy analysis.
Results
In CPRD, there were 127,826 eligible pregnancies indicated for antidepressants and 24,049 of those were exposed to antidepressants in trimester one. Among depressed patients, the risk of miscarriage for antidepressant users was slightly elevated in the adjusted model (aHR 1.06 95%CI 1.01–1.09). For anxiety and other indications, the confidence intervals crossed the null. For all indications, there was no evidence of an association between antidepressant use during trimester one and stillbirth. When comparing prevalent and incident users in trimester one, incident use appeared to be associated with a higher risk of miscarriage for depression and anxiety (aHR 1.23 95%CI 1.10–1.38 and aHR 1.25 95%CI 1.08–1.45, respectively) than prevalent use. Results from the additional analyses will be presented.
Conclusions
These early findings suggest that the risk of outcomes may be differential based on whether women were new initiators or on an established antidepressant regimen pre-pregnancy. The causal nature and meaning of these observations will be discussed.
mostly in the first few weeks. Although these medications are not advised against during pregnancy for those who may need them, it is still unknown whether they causally increase the risk of adverse pregnancy outcomes, such as miscarriage or stillbirth. Whether these observed associations are causal, or an artefact of bias is yet to be confirmed. We aimed to use the Clinical Practice Research Datalink (CPRD), a UK-wide repository of primary care data, and a range of comparisons to investigate antidepressant use during trimester one and pregnancy loss.
Methods
We included pregnancies that occurred between 1996 and 2018 and among women who had been diagnosed with an indication for antidepressants (depression, anxiety or other (including migraine prophylaxis and stress incontinence). Exclusions were made based on outcome and follow-up time. Antidepressant use during trimester one was defined as having at least one prescription overlapping with or starting between the start date of pregnancy and the start of trimester two. Exposed pregnancies were also split into prevalent users (those using antidepressants before pregnancy) and incident users (those not using antidepressants before pregnancy) for a secondary analysis. Miscarriage and stillbirth were defined in the Pregnancy Register, supplemented with Hospital Episode Statistics (HES) where possible. Hazard ratios and 95% confidence intervals were generated using multivariable Cox proportional hazards models. Additional analyses are ongoing, including a propensity score analysis and exposure discordant pregnancy analysis.
Results
In CPRD, there were 127,826 eligible pregnancies indicated for antidepressants and 24,049 of those were exposed to antidepressants in trimester one. Among depressed patients, the risk of miscarriage for antidepressant users was slightly elevated in the adjusted model (aHR 1.06 95%CI 1.01–1.09). For anxiety and other indications, the confidence intervals crossed the null. For all indications, there was no evidence of an association between antidepressant use during trimester one and stillbirth. When comparing prevalent and incident users in trimester one, incident use appeared to be associated with a higher risk of miscarriage for depression and anxiety (aHR 1.23 95%CI 1.10–1.38 and aHR 1.25 95%CI 1.08–1.45, respectively) than prevalent use. Results from the additional analyses will be presented.
Conclusions
These early findings suggest that the risk of outcomes may be differential based on whether women were new initiators or on an established antidepressant regimen pre-pregnancy. The causal nature and meaning of these observations will be discussed.
| Original language | English |
|---|---|
| Pages | 47 |
| Number of pages | 1 |
| Publication status | Accepted/In press - 17 May 2023 |
| Event | British Association of Psychopharmacology Summer Meeting 2023 - University of Manchester, Manchester, United Kingdom Duration: 23 Jul 2023 → 26 Jul 2023 https://www.bap.org.uk/BAP2023 |
Conference
| Conference | British Association of Psychopharmacology Summer Meeting 2023 |
|---|---|
| Abbreviated title | BAP 2023 |
| Country/Territory | United Kingdom |
| City | Manchester |
| Period | 23/07/23 → 26/07/23 |
| Internet address |