Vitamin K antagonists (VKA) are known to act as teratogens; however, there is still uncertainty about the relative risk for birth defects and the most sensitive period. In a multi-centre (n = 12), observational, prospective study we compared 666 pregnant women exposed to phenprocoumon (n = 280), acenocoumarol (n = 226), fluindione (n = 99), warfarin (n = 63) and phenindione (n = 2) to a non-exposed control group (n = 1,094). Data were collected by institutes collaborating in the European Network of Teratology Information Services (ENTIS) during individual risk counselling between 1988 and 2004. Main outcome measures were coumarin embryopathy and other birth defects, miscarriage rate, birth-weight, and prematurity. The rate of major birth defects after 1st trimester exposure was significantly increased (OR 3.86, 95% CI 1.86-8.00). However, there were only two coumarin embryopathies (0.6%; both phenprocoumon). Prematurity was more frequent (16.0% vs. 7.6%, OR 2.61, 95% CI 1.76-3.86), mean gestational age at delivery (37.9 vs.39.4, p<0.001), and mean birth weight of term infants (3,166 g vs. 3,411 g; p < 0.001) were lower compared to the controls. Using the methodology of survival analysis, miscarriage rate reached 42% vs. 14% (hazard ratio 3.36; 95% CI 2.28-4.93). In conclusion, use of VKA during pregnancy increases the risk of structural defects and other adverse pregnancy outcomes. The risk for coumarin embryopathy is, however, very small, in particular when therapy during the 1(st) trimester did not take place later than week 8 after the 1(st) day of the last menstrual period. Therefore, elective termination of a wanted pregnancy is not recommended if (inadvertent) exposure took place in early pregnancy. Close follow-up by the obstetrician including level II ultrasound should be recommended in any case of VKA exposure during pregnancy.