Pregnant women with mechanical heart valves are an ultrahigh-risk group, in whom all anticoagulant options pose maternal and/or fetal risks. It is unclear why the authors of the recent update 1 selected 4 studies, neither the most recent nor of the highest quality, to give a warfarin embryopathy rate of 1.6%. This anomaly rate is not supported by evidence in their Table 1 (3.9% of all pregnancies and 7.4% of livebirths) or by a recent systematic review (6.4% increasing to 10.2% in prospective

more »

... udies). 2 Furthermore, in the Vitale et al 3 study, which reported the low anomaly rate with warfarin doses of Ͻ5 mg, women in 43% of pregnancies required Ͼ5 mg warfarin, and among these 25 pregnancies, only 6 babies survived; of these 6 babies, 3 had anomalies. The clinician should be cognizant of the high rate of late fetal losses and neonatal deaths associated with warfarin, an issue of particular concern to women. In the systematic review, late fetal losses or neonatal deaths in warfarin-exposed pregnancies occurred in 8.8% of pregnancies and 11.7% of births, 10-fold that expected in a general population. 2 A similar rate of 13.6% was found in 3 studies (nϭ103) published after the final date for inclusion in that review. Consequently, despite warfarin being associated with a lower rate of pregnancy-associated valve thrombosis, some women are reluctant to continue taking warfarin while pregnant. Experience with low molecular weight heparin in pregnant women with mechanical valves is limited, and there are reports of case fatalities. We published a prospective case series of 14 pregnancies in women with mechanical valves treated with therapeutic enoxaparin (1 mg/kg twice daily). 4 One nonfatal valve thrombosis occurred at 20 weeks' gestation. Hung and Rahimtoola 1 comment on the recent enoxaparin product warnings, pointing out, "This product . . . is not recommended for thrombotic prophylaxis in patients with prosthetic heart valves" (p 1241). In response to the restrictive labeling of enoxaparin, a recent cardiology consensus statement concluded that the role of enoxaparin in pregnant women with mechanical valves "should more appropriately be considered an unproven and imperfectly studied alternative among a trio of suboptimal and potentially unfavorable options (warfarin, [unfractionated heparin], low molecular weight heparins)" (reference 5, p 16). Further, as enoxaparin does not cross the placenta, teratogenic effects are implausible. In our service, pregnant women with mechanical heart valves are provided with written information identifying the risks and benefits for all anticoagulant options. After the woman agrees to use an anticoagulant regimen, written consent is obtained. We agree with the consensus statement that "concern about efficacy and safety was justified for all agents" (p 16) and "the selective warning language about enoxaparin in pregnant patients with mechanical valves appears not to be justified" (reference 5, p 15). Finally, in our series of 41 women with homografts who became pregnant, 94% of the pregnancies resulted in a livebirth, and there were no cases of structural valve deterioration. 6