Safety of direct oral anticoagulant exposure during pregnancy: a retrospective cohort study

Background Direct oral anticoagulants (DOACs) have largely replaced vitamin K antagonists in many indications for anticoagulation. Prescribed to millions of patients, including women of reproductive age, exposure to DOACs in early pregnancy is not uncommon, but data on the embryotoxic risks are scarce. We aimed to assess the risk of DOAC embryotoxicity in a large sample of reported cases. Methods In this retrospective cohort study, we collected individual case reports of DOAC exposure in pregnancy from gynaecologists, haematologists, and vascular specialists starting from May, 2015. We obtained exports in April and October, 2017, August, 2018, and December, 2019, from the pharmacovigilance databases of the DOAC manufacturers, the European Medicines Agency (EMA), the German drug authority, and searched the homepage of the US Food and Drug Administration (FDA) for pregnancy exposure reports. Data from the International Society of Thrombosis and Haemostasis (ISTH) registry were obtained in August, 2018, and on July 21, 2020; data from the Teratology Information Service in Ulm, Germany, were received July 22, 2020. We also ran a systematic literature search on July 22, 2020, for cases of DOAC exposure. These data were compiled with those from our 2016 risk assessment and duplicate reports were excluded. Fetal or neonatal abnormalities were classified as a major birth defect according to the European Concerted Action on Congenital Anomalies and Twins (EUROCAT) classification and adjudicated into four categories: relation to DOAC exposure likely, possible, unlikely, or unrelated. Findings We identified 1193 reports of DOAC exposure during pregnancy: 49 from physicians, 48 from the ISTH registry, 29 from the Teratology Information Service, 62 from the German drug authority, 536 from Bayer (extracted from the Bayer pharmacovigilance system, the WHO VigiBase, and from the FDA Adverse Event Reporting System), 87 from Boehringer Ingelheim, 16 from Daiichi Sankyo, 98 from the literature search, two from the FDA homepage search, ten from the Risk Evaluation and Mitigation Strategy Review, and 256 from the EMA reports. After excluding potential duplicates, we identified 614 unique cases of DOAC exposure in pregnancy occurring between Feb 1, 2007, and July 9, 2020, that consisted of rivaroxaban in 505 (82%) pregnancies, dabigatran in 36 (6%), apixaban in 50 (8%), and edoxaban in 23 (4%). The median duration of DOAC exposure was 5·3 weeks (IQR 4·0–7·0) into pregnancy. Information on pregnancy outcome was available in 336 (55%) of 614 pregnancies: 188 (56%) livebirths, 74 (22%) miscarriages, and 74 (22%) elective pregnancy terminations. 21 (6%; 95% CI 4–9) of 336 showed fetal abnormalities, of which 12 (4%; 2–6) were adjudicated as major birth defects potentially related to DOAC exposure. Interpretation Although reports of pregnancy outcomes after DOAC exposure are missing important details and predominantly describe rivaroxaban exposures, the available data do not suggest that DOAC exposure in pregnancy carries a high risk of embryopathy. The 2016 ISTH guidance against elective pregnancy termination for fear of DOAC embryotoxicity and the recommendation in favour of close pregnancy surveillance is still valid. Pregnancy outcome data are inconsistently captured in pharmacovigilance databases, indicating a strong need for a more robust system of reporting. Funding None.