The fetal circulation and essential organs—a new twist to an old tale

The examination of the human fetal cardiovascular system is of central interest to our understanding of normal fetal development, as well as to the evaluation of fetal disease. For a long time, the circulation of the human fetus was explored predominantly on the arterial side and interpreted based on animal experimental data. With the more recent expansion of examination techniques into the venous system and the incorporation of blood-flow volume measurements, some traditional concepts on the parallel functionality of the fetal circulation may be challenged. Although the fetal circulation serves as a parallel conduit for bloodstreams with different nutritional content, the partitioning of nutrient flux between the four cardinal organs (liver, heart, brain and placenta) occurs sequentially. This serial dimension of the fetal circulation is best appreciated when reviewing its functional anatomy, the dynamics of the primary shunts and watershed areas, and the adaptations that have been documented in fetal health and disease. Blood with the highest concentration of oxygen and substrates enters the fetus via the umbilical vein and reaches the liver as the first major organ. The ductus venosus is the first shunt that determines the proportional distribution of these nutrients between the liver and the central circulation. The watershed area associated with the ductal shunt is the left portal vein, where umbilical venous blood that continues to the liver comes in contact with depleted portal blood that drains the splanchnic circulation1. The heart is the next major organ receiving bloodstreams with a range of nutritional contents from different sources. Among the right atrial tributaries, the ductus venosus and left hepatic vein carry blood with higher nutritional content than do the other venous tributaries (inferior and superior venae cavae, right and middle hepatic veins and the coronary sinus). On the left side, the pulmonary veins return depleted blood to the left atrium. The foramen ovale is the second shunt partitioning these incoming bloodstreams. Due to their different directionality and velocities, the position of the crista dividens and valve of the foramen ovale, saturated blood from the ductus venosus reaches the left ventricle preferentially, while the relatively depleted blood enters the right ventricle. The preductal aorta delivers nutrient-rich blood to the myocardium and brain (via the brachiocephalic circulation), while less saturated blood from the right ventricle reaches the lungs and the ductus arteriosus. The ductus arteriosus serves as a conduit that unites these two bloodstreams through its insertion into the aorta distal to the left subclavian artery. The aortic isthmus is the associated watershed area where shunting between bloodstreams originating from the left and right ventricles occurs. Downstream of the ductus arteriosus, the descending aorta carries blood with a nutritional content that results from the mixture of these two bloodstreams. The umbilical arteries provide the fourth shunt where depleted blood is channeled to the placenta for gas, nutrient and fluid exchange (Figure 1). Diagram illustrating the serial partitioning of nutrient- and oxygen-rich blood reaching the fetus via the umbilical vein. The first partition at the level of the ductus venosus distributes the majority of umbilical venous blood to the liver. Umbilical venous blood that continues towards the heart is partitioned towards the left ventricle (LV) at the foramen ovale. This blood supplies the brain and upper part of the body via the brachiocephalic circulation and the myocardium via the coronary circulation. A minor proportion of blood from the right ventricle (RV) supplies the lungs, while the remainder continues through the ductus arteriosus (DA) towards the aorta. At the aortic isthmus, bloodstreams directed towards the descending aorta are partitioned based on the relationship of blood-flow resistance in the brachiocephalic and subdiaphragmatic circulations. While net forward flow is maintained under physiological conditions, diastolic flow reversal occurs when brachiocephalic resistance falls and/or subdiaphragmatic (placental) resistance rises. Finally, the major proportion of descending aortic blood is partitioned at the umbilical arteries, to return to the placenta for respiratory and nutrient exchange. The dynamics of the ductus venosus shunt impact the fetal circulation in several ways. The smaller diameter of the ductus venosus in relation to the umbilical vein and its trumpet shape achieve significant blood-stream acceleration, so that forward delivery of blood against the cardiac pressure gradient can be achieved. The generation of this high-velocity jet is critical to achieve intracardiac separation of bloodstreams, and thus also impacts downstream partitioning of blood flow at the level of the foramen ovale. Changes in ductus venosus diameter, by an as yet undefined mechanism, regulate the partition of umbilical venous blood between the liver and the heart: dilation results in a greater diversion to the heart, while narrowing channels more blood towards the liver. The overall forward delivery of blood towards the heart is counterbalanced by the ability of the heart to accommodate venous return. This in turn depends on cardiac compliance, contractility and afterload. In the human fetus, the fraction of ductus venosus shunting decreases as pregnancy advances. At 18–20 weeks' gestation, up to 32% of umbilical venous blood bypasses the liver. Near term, 18–25% of umbilical venous flow shunts through the ductus venosus to reach the right atrium in this high-velocity stream, while 55% reaches the dominant left hepatic lobe and 20% the right liver lobes2, 3. At the same time, improved efficiency of forward cardiac function is reflected in a progressive decline in venous Doppler indices with advancing gestation4. Due to its capacity for dynamic change the ductus venosus behaves functionally like an arterial vessel. Shunting through the foramen ovale is critical for the delivery of enriched blood to the myocardium and brain. Shunt dynamics are affected by the jet of blood coming in from the ductus venosus and downstream blood-flow resistance of the right and left ventricles. Right ventricular afterload is determined mainly by the pulmonary arteries, ductus arteriosus, subdiaphragmatic circulation and the placenta, while left ventricular afterload depends primarily on brachiocephalic and coronary impedance. In utero, the right ventricle receives a greater proportion of venous return, and also contributes to a greater proportion of combined cardiac output (typically 60%). Eleven percent of cardiac output is distributed to the lungs and 46% reaches the ductus arteriosus. In the second trimester, the foramen ovale transmits approximately 33% of the combined cardiac output to the left side and contributes 76% of the left ventricular filling5. With advancing gestation, this right-to-left shunting across the foramen ovale decreases by 45% due to a corresponding increase in pulmonary blood flow. Accordingly, the foramen ovale only contributes 50% of the left ventricular filling by 30 weeks' gestation. Advancing gestation is therefore associated with increasing diversion of cardiac output to the lungs, with a concomitant increase of pulmonary venous return to the left atrium6. Although the ductus arteriosus is traditionally described as a shunt between the right and left ventricular outputs, it is the aortic isthmus where this shunting occurs7. Located between the left subclavian artery and the aortic insertion of the ductus arteriosus, the direction of diastolic blood flow in the aortic isthmus is determined by the relationship between vascular impedance in the subdiaphragmatic and brachiocephalic circulations. If blood-flow resistance in the subdiaphragmatic circulation is lower than that in the brachiocephalic circulation, blood flow is forward throughout the cardiac cycle. If the converse is true, diastolic blood flow in the aortic isthmus may reverse. With the physiological decrease in brachiocephalic blood-flow resistance with advancing gestation, early diastolic reversal or aortic isthmic flow may be observed from 25 weeks onwards. While this trend becomes more marked towards term, the net flow (systolic forward flow/diastolic reversal) is always antegrade under physiological conditions. Therefore, advancing gestation is associated with a successive increase in the right-to-left shunt at this level. The umbilical arteries are the only means by which depleted blood coming from the fetus can reach the placenta. The control of this fundamental vascular connection, if any, has not been a major research focus. The partition of blood flow at this level can be altered by changes in blood-flow resistance in the placenta and fetal circulation in the pelvis and lower limb. Typically, the lower blood-flow resistance in the umbilical arteries favors partition of blood flow towards the placenta. Elevation of blood-flow resistance in the arteries of the limbs will augment this diversion of blood flow towards the placenta, while elevation in placental blood-flow resistance will have the opposite effect. During normal fetal development, the distribution of nutrient-rich blood from the placenta follows a characteristic pattern. Initially, a larger proportion of blood bypasses the liver. This occurs at a time when the placental growth velocity precedes exponential fetal growth8. When fetal growth accelerates, the hepatic diversion of nutrients increases and it has been suggested that this increase in ‘hepatic dominance’ may be instrumental in fetal growth regulation9. At the cardiac level the contribution of nutrient-rich venous blood to left ventricular filling through the foramen ovale decreases as pulmonary blood flow and venous return increase. At the same time, progressive diastolic shunting towards the brachiocephalic circulation is observed in the aortic isthmus. At the level of the umbilical arteries, a progressive decline in placental blood-flow resistance maintains perfusion towards the placenta. Assessment of these distributional dynamics becomes most relevant in conditions under which fetal nutrient supply is limited. Redistribution has been defined as a shift in the proportional contribution of individual ventricles to the total cardiac output10. When the relative contribution of the left ventricle rises, a higher proportion of nutrient-rich blood is channeled towards the myocardium and the upper body and brain through the brachiocephalic circulation. This concept is an oversimplification. Firstly, the physiological distribution of nutrients from the placenta places essential organs in a serial order (liver, myocardium, brain and placenta). Secondly, partitioning of nutrients also occurs at successive levels in the circulation (ductus venosus, foramen ovale, aortic isthmus, umbilical arteries). Finally, individual essential organs serve different roles in the human fetus and autoregulatory adjustments of organ perfusion have been reported for the liver11, heart12, and brain13, and postulated for the placenta14. These organ-sparing effects are also invoked in a sequential manner with advancing fetal compromise. It therefore appears worth considering the fetal circulation as a serial unit distinguishing between venous redistribution, arterial redistribution and individual organ-sparing effects. The shunt dynamics in the ductus venosus–left portal vein system are responsive to umbilical venous nutritional content, decreased umbilical venous volume flow and marked elevations in placental blood-flow resistance15-18. An unbalanced maternal diet in women with low body-fat content is associated with a decrease in ductus venosus diameter and increased umbilical venous diversion to the liver. On the other hand, a reduction in umbilical venous volume flow results in ductal dilation and increased umbilical venous diversion, maintaining overall ductus venosus contribution to the heart. The exact mechanisms that regulate these adjustments in ductus venosus diameter are under investigation19. With increased ductus venosus shunting towards the heart, marked reversal of flow in the left portal vein towards the ductus venosus may be observed. This suggests that the ductus venosus may receive a variable admixture of depleted splanchnic blood due to left portal shunting. To summarize, venous redistribution potentially affects the nutritional supply of all downstream organs. If diversion to the liver is increased, ductus venosus contribution of nutrient-rich blood to the left ventricle decreases. If diversion towards the heart increases, volume flow may be maintained, or increased. The proportion of portal shunting towards the ductus potentially alters the nutritional content of this bloodstream. Assessment for venous redistribution could therefore prove critical in the study of fetal programming and to refine our assessment of arterial redistribution. Arterial redistribution occurs at the level of the foramen ovale and the aortic isthmus. An increase in pulmonary vascular resistance, as may be observed in fetal growth restriction20, increases intracardiac right-to-left shunting through the foramen ovale into the left ventricle. Therefore, there is a rise in the contribution of blood with higher nutritional content from the ductus venosus to left ventricular filling. Elevation of blood-flow resistance in subdiaphragmatic and placental vascular beds or a decline in brachiocephalic impedance impact at a cardiac level and in the aortic isthmus. This relative rise in right ventricular afterload promotes right-to-left shunting through the foramen ovale. At the aortic isthmus, diastolic reversal of blood flow indicates increased diastolic diversion of blood originating from the right ventricle towards the brachiocephalic circulation. The overall effect of these central adjustments in shunting is additive. The relative increase in left ventricular output enhances myocardial and brachiocephalic perfusion, while aortic isthmus shunting supplements cephalic diversion of blood as long as the net flow of systolic and diastolic forward velocities is forward21. When the net flow in the aortic isthmus becomes retrograde, the central shift of cardiac output towards the left ventricle is no longer observed. It has been suggested that the nutrient and oxygen content of the left ventricle drops under these circumstances; this is supported by the increased risk for adverse childhood neurodevelopment in fetuses that exhibit net reversal of aortic isthmus blood flow22. Distal to the confluence of the ductus arteriosus and aorta, ongoing perfusion of the placenta is critical to maintain adequate nutrient exchange. Marked elevation in blood-flow resistance in the pelvis and lower extremities favors diversion of blood flow towards the umbilical vessels. This mechanism may play a role in preserving placental perfusion in severe placental insufficiency23. Venous and arterial redistribution can only be effective as long as adequate forward cardiac function can be maintained. In severe placental insufficiency with marked elevations of placental afterload and/or myocardial dysfunction, central venous pressure becomes elevated as forward cardiac function declines24. Facilitated retrograde transmission of the atrial pressure pulse through a dilated ductus venosus, with a subsequent decline in the a-wave, is observed under these circumstances. The contribution of the ductus venosus to left ventricular filling may decrease under these conditions. On the arterial side of the circulation, forward cardiac function may become insufficient to maintain net antegrade flow in the aortic isthmus. Under these circumstances, intracardiac right-to-left shunting becomes inefficient and arterial redistribution does not occur. Individual organ-sparing effects may therefore be invoked at any time when local organ perfusion becomes insufficient. Local autoregulatory vascular adjustments that control organ perfusion are of three principal types. The first type acts synergistically with venous and arterial redistribution to augment organ blood flow. The second type counteracts an ‘organ-steal effect’ that may occur as a result of redistribution. The third type is relatively independent of redistribution, because the vascular beds supplying the organs arise distal to the ductus arteriosus and therefore are little affected by redistribution. Of the local organ-sparing effects, brain sparing is the longest recognized in the human fetus, while heart, liver and adrenal sparing25 were described more recently. The heart is the driving force of the fetal circulation and also has a unique metabolic role, since myocardium can metabolize lactate and may remove up to 80% of circulating lactate under certain conditions26. The brain has similar metabolic capacities, as it can switch its major fuel source from glucose to ketone bodies, which is advantageous during fetal starvation27. Both of these organs derive their blood supply from the left ventricle. Venous redistribution towards the heart and arterial redistribution through the foramen ovale and aortic isthmus lead to increased diversion of blood towards these organs. Vasodilatation at the organ level acts synergistically to increase organ blood flow. Heart sparing may be acute, chronic, or acute-on-chronic. Acute coronary vasodilatation in response to myocardial ischemia may recruit a myocardial flow reserve of approximately four times the basal flow. Chronic heart sparing is due primarily to redistribution. Acute-on-chronic heart sparing occurs with advanced cardiovascular deterioration. As chronic hypoxia induces coronary angiogenesis, massive augmentation of the coronary blood flow (up to 12-fold greater than the basal flow rate) can be triggered if acute myocardial hypoxia supervenes. Brain sparing as a result of cerebral vasodilatation results in a decrease in the Doppler index. However, similar Doppler changes are observed with the development of fetal hypertension28. Concurrent examination of the aortic isthmus may be helpful to refine the assessment of the cerebral arterial Doppler changes. The liver is unique because it derives its blood supply from venous and arterial sources. These include the umbilical vein, the splanchnic portal system and the hepatic artery. In prolonged venous redistribution, diversion of umbilical venous blood away from the liver sets up a state of chronic hepatic insufficiency. When placental insufficiency deteriorates further and baseline arterial supply becomes insufficient, hepatic artery vasodilatation is invoked to recruit additional hepatic blood-flow reserve from the arterial circulation. This vascular state is typically associated with hepatic dysfunction in the fetus and a marked deterioration of the acid–base balance29. Liver sparing is therefore unique, since, unlike brain and heart sparing, it does not act synergistically with venous and arterial redistribution. Because the blood supply to the adrenal glands arises after the confluence of the ductus arteriosus and the aorta, redistribution should have little impact on overall adrenal perfusion. It is therefore more likely that the augmentation of blood flow in this organ is due primarily to local autoregulation. Prior observations made in the animal model are increasingly being replaced by human data. However, because our ability for simultaneous assessment of several fetal vascular beds is limited, we still have to rely on animal experiment data. Experience from the sheep model verifies that decreased oxygen delivery through the umbilical vein results in increased diversion of cardiac output to the brain, heart and adrenal glands at the expense of the carcass, skin and scalp. This is achieved through venous and arterial redistribution as described above, and most likely autoregulatory mechanisms that have not been evaluated concurrently. In this setting, adrenal artery blood flow increases by almost 300%, while myocardial and brain blood flow increase by 120% and 80%, respectively30. Between the fetal liver and the ductal insertion into the aorta, the fetal circulation consists of a series of shunts that achieve sequential partitioning of bloodstreams with different nutritional content. Active partitioning of nutrient-rich umbilical venous blood by the ductus venosus affects global downstream distribution of nutrients. Passive shunting through the foramen ovale and aortic isthmus predominantly affects nutrient partitioning between the heart, brain and lungs. In contrast to the venous side of the circulation, arterial distribution is primarily redundant due to differential effects of afterload on foramen ovale and aortic isthmus shunting. Deviations from the physiological states can result in venous and arterial redistribution. Heart and brain sparing act synergistically with redistribution to augment organ flow. In contrast, liver sparing is invoked when excessive venous redistribution jeopardizes hepatic supply. Beyond the ductus arteriosus, adjustments of organ flows may be less dependent on redistribution, relying on organ autoregulation to recruit blood-flow reserve.