Anticoagulation for pregnant women with mechanical heart valves: a systematic review and meta-analysis

Aims To review maternal and foetal outcomes in women with mechanical heart valves (MHVs) treated with vitamin-K antagonists (VKAs), first-trimester heparin followed by VKAs (sequential treatment), low molecular weight heparin (LMWH) and unfractionated heparin (UFH) during pregnancy, in order to inform practice. Methods and results Medline, Embase and Central were searched from inception until February 2016. Two reviewers independently screened 1786 titles, reviewed 110 full-texts and extracted data and assessed risk-of-bias from 46 articles. Pooled incidence (95% confidence intervals) was calculated for maternal and foetal outcomes. Included studies had a moderate or high risk-of-bias. With VKAs, sequential treatment and LMWH, maternal mortality occurred in 0.9% (0.4-1.4), 2.0% (0.8-3.1) and 2.9% (0.2-5.7), thromboembolic complications in 2.7% (1.4-4.0), 5.8% (3.8-7.7) and 8.7% (3.9-13.4), livebirths in 64.5% (48.8-80.2), 79.9% (74.3-85.6) and 92.0% (86.1-98.0) and anticoagulant-related foetal/neonatal adverse events (embryopathy or foetopathy) in 2.0% (0.3-3.7), 1.4% (0.3-2.5) and 0%, respectively. When UFH is used throughout pregnancy, 11.2% (2.8-19.6) suffered thromboembolic complications. Foetal loss and adverse events occurred with first-trimester warfarin doses ≤ 5 mg/day, although there were more livebirths [83.6% (75.8-91.4) vs. 43.9% (32.8-55.0)] and fewer foetal anomalies [2.3% (0.7-4.0) vs. 12.4% (3.3-21.6)] with lower doses than with warfarin > 5 mg/day. Conclusions VKAs are associated with fewest maternal complications but also with fewest livebirths. Sequential treatment does not eliminate anticoagulant-related foetal/neonatal adverse events. LMWH is associated with the highest number of livebirths. The safety of UFH throughout pregnancy and first-trimester warfarin  ≤ 5 mg/day remains unconfirmed.