Seventy-eight women at earlier than 35 weeks' gestation with premature rupture of membranes and/or preterm labor were randomly assigned to receive either 10 mg vitamin K1 intramuscularly (IM) or no treatment. If delivery did not occur within 4 days, the dose of vitamin K1 was repeated. Women whose pregnancies continued beyond 8 days received 20 mg of vitamin K1 orally every day until the end of the 34th week or until delivery, whichever occurred earlier. The median maternal plasma vitamin K1 level was significantly higher in treated than in untreated subjects (11.592 versus 0.102 ng/mL; P less than .001). The median cord plasma levels were 0.024 ng/mL in the treated group and 0.010 ng/mL in the controls, a significant difference (P = .046). Median plasma vitamin K1 levels were comparable in mothers receiving the drug by the IM route only and by both the IM and oral routes (10.533 versus 11.928 ng/mL; P = .460). The infants of the latter group, however, had significantly higher median cord plasma levels (0.42 versus 0.017 ng/mL; P less than .001). There was no correlation between cord plasma vitamin K1 levels and gestational age or duration of maternal supplementation with vitamin K1. We conclude that, in preterm pregnancies, vitamin K1 crosses the placenta slowly and to a limited degree.