Fetal Liver-Sparing Cardiovascular Adaptations Linked to Mother’s Slimness and Diet
Fetal adaptations to impaired maternoplacental nutrient supply include altered regional blood flow. Whether such responses operate within the normal range of maternal body composition or diet is unknown, but any change in fetal liver perfusion could alter hepatic development, with long-term consequences for the risk of cardiovascular and metabolic disease. In 381 low-risk pregnancies, we found that the fetuses of slimmer mothers with lower body fat stores and those eating an unbalanced diet had greater liver blood flow and shunted less blood away from the liver through the ductus venosus at 36 weeks gestation. Consequences of such “liver-sparing” may underlie the increased cardiovascular risk of people whose mothers were slimmer and had lower body fat stores in pregnancy.
The processes controlling the growth of individual fetal organs are incompletely understood, but during mammalian development, changes in blood flow to an organ affect its growth and can have major lifelong effects on organ function. The fetus uses changes in blood flow to defend itself against an insult such as reduced oxygen availability. By distributing blood pumped by the heart toward essential organs, eg, the brain (“brain-sparing”), at the expense of organs such as muscle, gut and liver, the fetus reduces nonessential growth and economizes on oxygen demand. It is not known whether similar adaptive mechanisms regulate blood flow and growth in relation to fetal nutrient supply within the normal range: if so, such mechanisms may be invoked by maternal factors such as body fat stores and diet, and might in part underlie the link between the intrauterine environment and health in adulthood.1
Materials and Methods
To examine the relation between maternal body fatness and diet and fetal hepatic blood flow at 36 weeks gestation, we approached a low-risk population of 410 healthy women with singleton pregnancies and no major fetal abnormality. A study of this size has 90% power to detect a correlation coefficient of 0.16 between maternal body fatness and fetal hepatic blood flow at the 5% level. The women’s heights, weights, skinfold thicknesses, mid-upper arm circumferences, and diet had been measured by trained research nurses before pregnancy in the Southampton Women’s Survey.2
We used Doppler ultrasound (Acuson Sequoia) to measure blood flow in the umbilical vein and ductus venosus; the ductus venosus shunts a proportion of well-oxygenated placental blood past the fetal liver, toward the heart and head (Figure 1). We measured internal vessel diameter (late-diastole) and time-averaged maximum velocity (TAMX) (insonation angle <30°) in the intraabdominal umbilical vein (straight portion, before hepatic parenchymal branches) and at the ductus venosus inlet. Umbilical vein TAMX was obtained during a 3- to 5-second period or, if flow was pulsatile, as the mean during three heart cycles. Ductus venosus TAMX was calculated as the mean during three heart cycles. Blood flow (Q) was calculated as Q=h · (D/2)2 · π · TAMX, where D=vessel diameter (mean of 5 to 10 measurements) and h=spatial blood velocity profile coefficient (umbilical vein=0.5; ductus venosus=0.7).3 Intraclass correlation coefficients (random-effects regression) to assess intraobserver variation were 0.97 and 0.96 for the umbilical vein and ductus venosus diameter, respectively. We obtained complete data in 381 subjects (93%) and derived percent shunting from the ratio ductus venosus flow/umbilical vein flow; the correlation coefficient between ductus venosus cross-sectional area and shunting was 0.74 (P<0.001). Liver blood flow was derived as umbilical vein flow−ductus venosus flow.
Gestational age was calculated using menstrual data (61%) or, when these were uncertain or discrepant with ultrasound assessments, fetal anthropometry in early pregnancy. The Local Research Ethics Committee approved the study, and the women gave written informed consent. Where necessary, variables were transformed using logarithms to satisfy statistical assumptions of normality. Correlations between maternal characteristics and fetal blood flow were calculated as Spearman’s correlation coefficients on the continuously distributed variables using Stata 7.0 (Stata Corp, Texas).
Results and Discussion
Median (10th to 90th centile) values (n=381) were 30 years (25 to 35 years) for maternal age, 36+1 weeks (35+2 to 37+2 weeks) for gestation at examination, 3485 g (2875 to 4160 g) for infant birthweight, 24.1% (12.0% to 42.0%) for ductus venosus shunting, and 201.7 (138.9 to 293.2), 48.8 (23.1 to 86.8), and 150.9 (87.3 to 239.4) mL/min for umbilical vein, ductus venosus, and liver blood flows, respectively. The fetuses of slim women with thin subscapular skinfold thicknesses had reduced ductus venosus flow (r=0.17, P=0.001) and shunting (Figure 2), but increased liver blood flow (Figure 2); maternal subscapular skinfold thickness was not, however, related to fetal umbilical flow (P=0.3). Thinner suprailiac skinfold thickness and lower estimated body fat stores (sum of skinfold thickness measurements at the subscapular, triceps, biceps, and suprailiac sites) were similarly associated with reduced ductus venosus flow (r=0.20, P=0.0001 and r=0.21, P<0.0001, respectively) and shunting (r=0.24, P<0.0001 and r=0.24, P<0.0001, respectively) and with increased liver blood flow (r=−0.15, P=0.003 and r=−0.13, P=0.01, respectively); lower maternal body mass index was associated with reduced ductus venosus flow (r=0.16, P=0.002) and shunting (r=0.16, P=0.001), but only weakly with increased liver blood flow (P=0.23). Maternal subscapular/triceps skinfold thickness ratio (a measure of central adiposity), was not associated with ductus venosus flow (P=0.55) but a low ratio was associated with reduced shunting (r=0.11, P=0.03) and increased liver blood flow (r=−0.15, P=0.005). Independent of maternal body fatness, shunting and liver blood flow were not associated with maternal forearm muscle area, height, smoking, age, and social class, or with the heart rate of the fetus (all P>0.05).
To assess each woman’s diet, we used a validated 100-item food frequency questionnaire. Such questionnaires can identify similar patterns of diet as weighed diet records.4 Principal component analysis is a robust statistical method for characterizing dietary patterns from food frequency questionnaire data,5 and we have used this to derive three components that most discriminate between maternal dietary intakes.2 The first component yielded a dietary pattern consistent with patterns of food recommended for a “healthy” diet2; fetuses of women who ate an “imprudent” diet (including high intakes of chips/crisps, sugar, confectionary, white bread, soft drinks, and red meat and low intakes of fruit/vegetables, rice/pasta, yogurt, and wholemeal bread) had reduced ductus venosus shunting (P=0.04) and increased liver blood flow (P=0.02). These associations were independent of those with maternal slimness. The second component characterized a pattern that describes a “Western diet” with additional high intakes of fruit/vegetables, and the third component a pattern combining high intakes of vegetarian foods, confectionery, and snack foods; the dietary patterns described by the second and third components were not related to either ductus venosus shunting (P=0.45 and P=0.89, respectively) or fetal liver blood flow (P=0.83 and P=0.95, respectively).
After recruitment in early pregnancy, four women in our sample subsequently developed preeclampsia and 16 pregnancy-induced hypertension; eight women delivered infants weighing less than 2500 g. The associations between maternal subscapular skinfold thickness and imprudent diet and fetal ductus venosus shunting and liver blood flow were similar in the full sample and in those who had uncomplicated pregnancies. Within the narrow gestation range in our study, shunting and fetal liver blood flow were not related to gestational age (P=0.29 and P=0.70, respectively), making it unlikely that bias in gestational age assignment had an important effect on our findings.
Our findings suggest that human fetuses make cardiovascular responses to differing nutrient availability in the normal range. In slim women with low central adiposity and those eating an imprudent diet, altered nutrient availability will change umbilical venous blood nutrient content; fetal compensation requires increased hepatic nutrient interconversion and increased liver blood flow. In addition, animal studies show reducing ductus venosus shunting and augmenting liver blood flow experimentally can increase the growth of many fetal organs, perhaps by increasing hepatic substrate and growth factor output.6 Liver-sparing adaptations contrast with the brain-sparing response to fetal hypoxemia, which reduces hepatic blood flow and increases ductus venosus shunting (Figure 1), leading to disproportionate fetal growth. This suggests differing roles of the ductus venosus in normal and growth-restricted fetuses. In keeping with this, we found no overall association between shunting and infant birthweight in these low-risk pregnancies (r=−0.04, P=0.44), even though umbilical vein, ductus venosus, and liver blood flows had positive associations with birthweight (r=0.50, 0.27 and 0.40, respectively; all P<0.0001). As expected, a lower maternal body mass index was associated with lower infant birthweight (r=0.18, P=0.0003), but maternal slimness and lower subscapular skinfold thickness had only a weak and nonsignificant association with birthweight (r=0.08, P=0.11). Taking account of birthweight had little effect on the associations between maternal subscapular skinfold thickness and either shunting or liver blood flow.
Further research is required to replicate our findings, to define if they have long term effects on offspring physiology, and to examine the relation between maternal nutritional state and fetal liver blood flow in populations with differing maternal body composition and diet. For example, many women in North America eat unbalanced low carbohydrate/high protein diets; our findings cannot be applied directly to this dietary pattern, but they suggest such dietary imbalance could have important effects on fetal well-being. Experimental studies in fetal sheep show that blood distribution to the liver and ductus venosus is particularly sensitive to changes in umbilical venous pressure, blood viscosity, and an active regulation of diameter of the entire ductus venosus, resulting in altered regional blood flow.3 The mechanisms regulating ductus venosus diameter and blood flow through the fetal liver are incompletely understood, but may include shear stress and nitric oxide in addition to differences in catecholamine responses between the two pathways6,7; further research is required to define these mechanisms.
Although beneficial in utero, fetal liver blood flow adaptations in late pregnancy may set the function of perivenous and periportal hepatocytes, resulting in a “thrifty” hepatic metabolism as reported in offspring of protein-restricted pregnant rats.8 But they may have longer-term detrimental consequences for lipid and clotting factor homeostasis in the face of an enriched postnatal diet.
Our observations offer insights into the circulatory tuning of the fetal liver in relation to maternal body composition and diet. The concept of liver-sparing could lead to new diagnostic measures to investigate how maternal slimness and unbalanced diet increase the risk of adult cardiovascular disease and type-2 diabetes in the offspring.1
The study was funded by the British Heart Foundation, the UK Medical Research Council, the Dunhill Medical Trust, and the Research Council of Norway. We thank the study participants and staff, and all the general practitioners and practice staff who have assisted with the Southampton Women’s Survey.
Original received September 7, 2004; revision received November 1, 2004; accepted November 23, 2004.
Godfrey KM. Maternal nutrition and fetal development: implications for fetal programming. In: Barker DJP, ed. Fetal Origins of Cardiovascular and Lung Disease. New York: National Institutes of Health 2000: 249–271.
Kiserud T. Fetal venous circulation. Fetal Matern Med Rev. 2003; 14: 57–95.
Hu FB, Rimm EB, Smith-Warner SA, Feskanich D, Stampfer MJ, Ascherio A, Sampson L, Willett WC. Reproducibility and validity of dietary patterns assessed with a food frequency questionnaire. Am J Clin Nutr. 1999; 69: 243–249.
Desai M, Crowther NJ, Ozanne SE, Lucas A, Hales CN. Adult glucose and lipid metabolism may be programmed during fetal life. Biochem Soc Trans. 1995; 23: 331–335.