Anticoagulation for mechanical heart valves during pregnancy is essential to prevent thromboembolic events. Each regimen has drawbacks with regard to maternal or fetal risk. Objectives This meta-analysis sought to estimate and compare the risk of adverse maternal and fetal outcomes in pregnant women with mechanical heart valves who received different methods of anticoagulation. Methods Studies were identified using a Medline search including all publications up to June 5, 2016. Study inclusion required reporting of maternal death, thromboembolism, and valve failure, and/or fetal spontaneous abortion, death, and congenital defects in pregnant women treated with any of the following: 1) a vitamin K antagonist (VKA) throughout pregnancy; 2) low-molecular-weight heparin (LMWH) throughout pregnancy; 3) LMWH for the first trimester, followed by a VKA (LMWH and VKA); or 4) unfractionated heparin for the first trimester, followed by a VKA (UFH and VKA). Results A total of 800 pregnancies from 18 publications were included. Composite maternal risk was lowest with VKA (5%), compared with LMWH (16%; ratio of averaged risk [RAR]: 3.2; 95% confidence interval [CI] : 1.5 to 7.5), LMWH and VKA (16%; RAR: 3.1; 95% CI: 1.2 to 7.5), or UFH and VKA (16%; RAR: 3.1; 95% CI: 1.5 to 7.1). Composite fetal risk was lowest with LMWH (13%; RAR: 0.3; 95% CI: 0.1 to 0.8), compared with VKA (39%), LMWH and VKA (23%), or UFH and VKA (34%). No significant difference in fetal risk was observed between women taking ≤5 mg daily warfarin and those with an LMWH regimen (RAR: 0.9; 95% CI: 0.3 to 2.4). Conclusions VKA treatment was associated with the lowest risk of adverse maternal outcomes, whereas the use of LMWH throughout pregnancy was associated with the lowest risk of adverse fetal outcomes. Fetal risk was similar between women taking ≤5 mg warfarin daily and women treated with LMWH.