Despite the incidence of heart disease during pregnancy falling to 1.5% over the last 25 years, when a cardiac operation is required the risk is obviously greater as two lives are at risk. The risk to the mother is now similar to that for nonpregnant female patients (3% overall) but the fetal mortality remains high (19%). Cardiac operation is ill advised except in extreme emergencies during the first two trimesters as the incidence of teratogenesis is high. During the third trimester, with improvements in the outcome for premature infants with modern neonatal intensive care, delivery of the child immediately before commencing cardiopulmonary bypass is a safe option. If this is inappropriate, high-flow, high-pressure, normothermic bypass for as brief a period as possible should be used. However, although it has theoretic advantages, the benefit of pulsatile perfusion is unproven. The fetal response to cardiopulmonary bypass is bradycardia thought to be due to hypoperfusion secondary to uterine contractions, and this dysrhythmia is reversible by increasing the perfusion rate. Fetal heart rate monitoring is therefore essential to allow these manipulations. The response of the fetoplacental unit is more complex, comprising two elements: an early vasoactive response is due to prostaglandin synthesis, whereas a more profound late acidosis appears to be related to a fetal stress response. Whether these responses can be modified by changes in our approach to cardiopulmonary bypass in pregnant women remains to be proven. Finally, uterine contractions occur in response to bypass, possibly due to a dilutional effect from the stabilizing influence of progesterone. Various techniques to modify this include the administration of progesterone, beta2-agonists, and intravenous alcohol, all with some effect. Uterine monitoring is essential to allow early control of these contractions as they are associated with significant fetal loss.