SEARCH

SEARCH BY CITATION

Keywords:

  • blood flow;
  • circulation;
  • Doppler;
  • fetus;
  • growth;
  • liver;
  • macrosomia

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

Objective

To determine the venous and arterial hemodynamics underlying macrosomic fetal growth.

Methods

Fifty-eight healthy women who previously had given birth to a large neonate were included in a prospective longitudinal study. Of these, 29 gave birth to neonates with birth weight ≥ 90th percentile and were included in the statistical analysis. Umbilical vein blood flow and Doppler measurements of the ductus venosus, left portal vein and the hepatic, splenic, superior mesenteric, cerebral and umbilical arteries were repeated at 3–5 examinations during the second half of pregnancy and compared with the corresponding reference values. Ultrasound biometry was used to estimate fetal weight.

Results

Umbilical blood flow increased faster in macrosomic fetuses, showed less blunting near term and was also significantly higher when normalized for estimated fetal weight (P < 0.0001). The portocaval perfusion pressure of the liver (expressed by the ductus venosus systolic blood velocity) and the left portal vein blood velocity (expressing umbilical venous distribution to the right liver lobe) were significantly higher. Systolic velocity was higher in the splenic, superior mesenteric, cerebral and umbilical arteries, while the pulsatility index was unaltered in the cerebral, hepatic, splenic and mesenteric arteries, but lower in the umbilical artery.

Conclusions

There is an augmented umbilical flow in macrosomic fetuses particularly near term, also when normalized for estimated fetal weight, providing increased liver perfusion, including the right liver lobe. Signs of increased vascular cross section and flow are also seen on the arterial side but not expressed in the pulsatility index of organs with prominent auto-regulation (i.e. brain, liver, spleen and gut). Copyright © 2011 ISUOG. Published by John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

Fetal growth is closely linked to the placental circulation. At mid-gestation 30% of the combined fetal cardiac output is directed to the placenta, a fraction that declines to 20% near term1. During placental compromise this decline is augmented and appears at earlier stages of pregnancy1, 2. The fetus adapts to such constraints by diverting an increased fraction of the umbilical return from the liver to ensure oxygen delivery to the heart and brain through the ductus venosus3. However, it has become increasingly clear that the fetal liver is an important determinant of fetal growth. In the human fetus, the liver receives 75–80% of the umbilical flow4, and in fetal lambs the umbilical venous (UV) perfusion of the liver regulates hepatocyte proliferation and production of insulin-like growth factors 1 and 2. In turn, this stimulates proliferation and growth in the rest of the fetal body5, 6. Studies of UV distribution in physiologic pregnancies and intrauterine growth restriction3, 7, 8 support the assumption that these mechanisms also operate in human fetuses. Data also suggest that the UV perfusion of the liver determines fat accretion with postnatal consequences9 while the arterial circulation of the fetal liver and gut is more a marker of auto-regulation10–12 and compensatory responses to placental compromise and fetal growth restriction13. The umbilicocaval pressure gradient (which is the same as the portocaval pressure gradient) drives venous liver perfusion. In this study we utilized measurements of the peak systolic blood velocity in the portocaval shunt ductus venosus as a direct representation of this pressure gradient14–17. Accordingly, we here hypothesize that nutritional availability and macrosomic growth are reflected in venous return from the placenta and its distributional pattern rather than in arterial flow velocities and impedance.

The aim of the present study was to carry out a longitudinal assessment of the venous and arterial hemodynamic development of macrosomic fetuses during the second half of pregnancy.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

Subjects

For this prospective longitudinal study 58 healthy, non-diabetic pregnant women who had previously given birth to a large neonate (birth weight ≥ 4200 g) were recruited. The Regional Committee for Medical Research Ethics approved the study protocol (REK Vest no 203.03), and all participants gave prior written consent. Gestational age was determined by fetal head circumference at the routine ultrasound scan at 17–20 weeks' gestation18 or first-trimester measurement of crown–rump length19. Twin pregnancies and those with chromosomal aberrations, malformations or maternal diabetes were excluded from the study. Birth weight, placental weight, sex, gestational age at and mode of delivery and 5-min Apgar score were noted. Birth weight ≥ 90th centile according to Norwegian sex-specific standards20 was considered eligible for inclusion in the analysis.

The reference population consisted of 161 women with low-risk pregnancies included in a longitudinal study. Their population characteristics21 and Doppler reference ranges10–12, 21 have been described previously.

Ultrasound measurements

The participants were scheduled for four examinations at 3–5-week intervals during the second half of pregnancy. During each 1-hour session the participants were examined using a 2–5-, 2–7- or 4–8-MHz transabdominal transducer (Voluson 730 Expert, GE Medical Systems, Kretz Ultrasound, Zipf, Austria) with the high-pass filter set to 70 Hz. The women were examined for no more than 60 min in a semi-recumbent position with a pillow underneath the knees. We videotaped one typical examination session and timed the use of different ultrasound modes. Pulsed Doppler time constituted 11%, color Doppler 40% and gray-scale ultrasound 33% of the total examination time, the remaining 16% being frozen-image time. The mechanical and thermal indices were for the most part kept at < 1.1 and < 0.9, respectively, but occasionally reached 1.9 and 1.5.

Blood flow velocities, peak systolic velocity and time-averaged maximum velocity (TAMXV) were measured using Doppler ultrasound in the middle cerebral, umbilical, hepatic, splenic and superior mesenteric arteries, the intra-abdominal umbilical and left portal veins and the ductus venosus. Fetal weight was estimated from biometry of the head, abdominal circumference and femur length22. Blood-flow velocities in the study group were measured using the same techniques as in the reference studies10–12. Briefly, the umbilical artery was assessed in a free-floating loop of the umbilical cord and the middle cerebral artery at the proximal part of the vessel23. The hepatic artery was insonated in an axial or sagittal view of the fetal abdomen and flow velocity assessed close to the ductus venosus10. The splenic artery was identified in an axial view at its origin from the celiac artery in front of the aorta and posterior to the stomach24. The superior mesenteric artery was identified as the second of the anterior unpaired arteries from the abdominal aorta, and again the sample volume was placed at the proximal part of the vessel25. UV flow was estimated from repeated measurements (at least three) of the inner diameter and flow velocity in the intra-abdominal umbilical vein26. Prenatally, the left portal vein connects the umbilical vein to the portal circulation. We utilized blood velocity in the left portal vein as a direct reflection of the UV distribution to the right liver lobe27. The Doppler insonation was aligned to the left portal vein with a small sample volume placed between the ductus venosus inlet and the junction with the main portal stem in order to measure the TAMXV28. The ductus venosus blood velocity (reflecting the portocaval pressure gradient17) was assessed in a sagittal or oblique section of the fetal abdomen15. We determined flow velocity (in the veins for 2–4 s and in the arteries for at least three heart cycles) and pulsatility index (PI) during fetal quiescence over at least three uniform cardiac cycles.

Statistical analysis

Standard deviation scores were calculated based on power-transformed mean and SD values for the reference population and then modeled against gestational age by multilevel regression. Multilevel t-test was used to assess differences. Lack of overlap of the 95% CI of means was considered a significant difference. Otherwise, P⩽0.05 was regarded as statistically significant. Statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA) and the MLWin program (MLWin; Centre for Multilevel Modelling, University of Bristol, UK).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

Of the 58 participants, 29 gave birth to a large neonate with birth weight ≥ 90th centile according to Norwegian standards adjusted for gestational age and sex20. These 29 participants had 111 sessions of examinations and the corresponding number of data sets were included in the analysis and compared with the reference population. In the study population 1/29, 3/29, 20/29 and 5/29 participants had two, three, four and five sets of examinations, respectively. Characteristics of the study and reference populations are given in Table 1. All the women were Caucasian and only one was a smoker. There were four operative deliveries; two emergency Cesarean sections (one for threatening asphyxia and one for obstructed labor) and one elective Cesarean section (because of obstetric history). One forceps delivery was performed because of obstructed second stage of labor. None of the neonates had an Apgar score of < 7 at 5 min.

Table 1. Characteristics of the study population and the reference population
CharacteristicStudy population (n = 29)Reference population (n = 161)P
  • Data expressed as median (range) or %.

  • *

    Mann–Whitney U-test.

  • Pearson chi-square test.

  • Parity grouped as 0, 1 and 2+ for statistical analysis.

Maternal age at inclusion (years)31 (22–42)29 (20–40)< 0.05*
Parity2 (1–7)1 (0–5)< 0.0001
Body mass index at first visit (kg/m2)25.2 (20.1–45.8)22.9 (18.1–40.8)< 0.05*
Maternal height (cm)170 (160–186)167 (150–183)NS*
Maternal weight gain (kg)15.2 (0–26.5)14 (0–55)NS*
Gestational age at delivery (weeks)41 (38.1–43.0)40.4 (35.4–42.6)< 0.05*
Birth weight (g)4550 (4035–5210)3700 (2260–4980)< 0.05*
Male gender3150.30.056

Reasons for missing measurements were suboptimal visualization due to maternal adiposity, unfavorable fetal position, fetal movements and time constraints (Table 2). UV flow and UV distribution to the right liver lobe (assessed by TAMXV in the left portal vein) were higher during the second half of pregnancy in the fetuses that became macrosomic (Figures 1a and c and Table 2). Even more striking was the finding that the UV return was approximately 35% higher when normalized for estimated fetal weight (mL/min/kg; Figure 1b, Table 2). Ductus venosus peak systolic velocity—reflecting umbilicocaval perfusion pressure—was marginally higher in the macrosomic fetuses (Figure 2 and Table 2), whereas the arterial flow velocities were all significantly increased except for that of the hepatic artery (Table 2). Arterial impedance (as reflected by PI) was unaltered in all vessels except for the umbilical artery, where it was significantly lower in the study population.

thumbnail image

Figure 1. Umbilical venous (UV) blood flow (a), normalized UV blood flow (b), and distribution of UV blood to the right liver lobe expressed by left portal vein time-averaged maximum velocity (TAMXV) (c) in macrosomic fetuses (_____) and in the reference population (equation image). Non-overlapping 95% CI for the means (thin lines) indicates significant differences. Z-score statistics: P < 0.0001 for all variables.

Download figure to PowerPoint

thumbnail image

Figure 2. Individual values of ductus venosus peak systolic velocity (DV-PSV) in macrosomic fetuses (○) with mean (equation image) and 95% CI of the mean (equation image) of these values shown, along with the mean (equation image), 95% CI of the mean (equation image) and 5th and 95th centiles (equation image) of the reference population, showing a significant difference between groups (P = 0.019).

Download figure to PowerPoint

Table 2. Flow parameters of the macrosomic fetuses compared with the reference population10
 Study populationReference population 
ParameterMean ± SE (95% CI)nMean ± SE (95% CI)nP*
  • 95% CI given for the mean value. n given for number of measurements.

  • *

    t-test.

  • CI, confidence interval; DV, ductus venosus; HA, hepatic artery; LPV, left portal vein; MCA, middle cerebral artery; PI, pulsatility index; PSV, peak systolic velocity; SA, splenic artery; SE, standard error; SMA, superior mesenteric artery; TAMXV, time-averaged maximum velocity; UA, umbilical artery; UV, umbilical vein.

UV flow (mL/min)1.21 ± 0.11 (0.99 to 1.43)920 ± 0.05 (−0.09 to 0.09)525< 0.0001
UV flow normalized (mL/min/kg)0.85 ± 0.11 (0.64 to 1.06)900 ± 0.05 (−0.09 to 0.09)511< 0.0001
LPV-TAMXV (cm/s)0.75 ± 0.12 (0.52 to 0.98)71− 0.02 ± 0.06 (−0.14 to 0.1)272< 0.0001
DV-PSV (cm/s)0.26 ± 0.1 (0.06 to 0.46)960 ± 0.04 (−0.08 to 0.09)5500.019
HA-PSV (cm/s)− 0.17 ± 0.14 (−0.45 to 0.1)550.01 ± 0.08 (−0.15 to 0.17)1620.265
HA-PI0.08 ± 0.13 (−0.18 to 0.33)55− 0.01 ± 0.08 (−0.16 to 0.14)1620.57
SMA-PSV (cm/s)0.35 ± 0.1 (0.15 to 0.55)950 ± 0.04 (−0.08 to 0.08)5430.002
SMA-PI0.02 ± 0.1 (−0.17 to 0.21)95− 0.01 ± 0.04 (−0.09 to 0.08)5430.8
SA-PSV (cm/s)0.29 ± 0.11 (0.08 to 0.5)920.02 ± 0.05 (−0.07 to 0.11)4890.019
SA-PI0.12 ± 0.1 (−0.08 to 0.31)94− 0.02 ± 0.05 (−0.11 to 0.07)4890.2
MCA-PSV (cm/s)0.38 ± 0.11 (0.17 to 0.58)950.03 ± 0.04 (−0.06 to 0.12)524< 0.0001
MCA-PI0.18 ± 0.1 (−0.02 to 0.38)950 ± 0.04 (−0.08 to 0.09)5190.094
UA-PSV (cm/s)0.33 ± 0.1 (0.14 to 0.53)101− 0.03 ± 0.04 (−0.11 to 0.06)5620.001
UA-PI− 0.49 ± 0.1 (−0.68 to − 0.29)1010.01 ± 0.04 (−0.07 to 0.09)562< 0.0001

The birth weight to placenta weight ratio was the same in the study and the reference group, mean 5.3 (95% CI, 5.1–5.5) vs. 5.3 (95% CI, 4.9–5.6), respectively.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

In this study of healthy non-diabetic women with fetuses that became macrosomic, we found notably higher UV flow, higher venous perfusion pressure of the fetal liver and enhanced umbilical blood distribution to the right liver lobe during the second half of pregnancy compared with a reference population. The augmented umbilical flow was also substantial when normalized for estimated fetal weight. In low-risk pregnancies fetal growth is undiminished during the last weeks of pregnancy29, suggesting a continued increase in nutrient availability that is not correspondingly supported by the UV flow, which had a reduced increment near term (Figure 1a)4, 30. Interestingly, in the macrosomic fetuses, in addition to having higher UV flow per kg than expected, the UV flow also continued to increase during the last weeks of pregnancy, supporting the concept that the hemodynamic effect of umbilical liver perfusion is a determinant of growth in its own right. The concept is supported by experimental studies in sheep and a recent similar study of macrosomic fetuses5, 6, 31.

Increased TAMXV in the left portal vein has been shown to correlate well with increased distribution of UV blood to the right liver lobe27. In contrast, in intrauterine growth restriction, the UV supply to the liver and the right liver lobe is reduced3, 8, 13, 26, and in extreme cases of placental compromise the right liver lobe receives exclusively portal venous and no UV blood8. Our finding of increased UV hepatic perfusion pressure and high umbilical blood-flow velocity directed to the right liver lobe is in line with another study of macrosomic growth31 showing an increased UV liver supply, including the right liver. Such changes may be related to lasting metabolic and homeostatic patterns, since to be born large for gestational age is a risk factor for developing obesity and related diseases later in life32–34. Our study population had a higher body mass index than did controls, which is a known risk factor for increased birth weight, but no relationship was demonstrated between maternal weight gain and birth weight35, possibly due to small sample size.

In pregnancies complicated by diabetes, altered maternal substrate levels and placental nutrient transport and metabolism are thought to be major factors underlying excessive fetal growth36. In such pregnancies the placental weight to birth weight ratio was high37. In our study the women were non-diabetic and the relationship between birth weight and placental weight was normal, suggesting that abnormal placental metabolism and transport were not the primary cause of extreme growth in our study population.

An increased flow velocity in all arteries except the hepatic artery reflects increased volume flow and organ size38. Increased flow velocity in the splenic and superior mesenteric arteries is in agreement with increased portal return from the splanchnic organs in macrosomic fetuses31. However, hepatic artery flow velocity and PI did not differ from those of the reference population, although we expected the liver to be larger in the macrosomic fetuses. This fits with the assumption that the hepatic artery buffer response (HABR) was not activated in macrosomic fetuses. The HABR is known in postnatal life as a powerful auto-regulatory system that is activated when portal sinusoid perfusion is reduced39. The mechanism has also been shown to operate during prenatal life, a period when portal liver perfusion is dominated by the umbilical flow10–12. We observed signs of increased portal perfusion pressure and flow in the macrosomic fetuses (Figures 1 and 2 and Table 2), supporting our assumption that the hepatic artery parameters reflect auto-regulatory status rather than actual liver size. A similar pattern of unaltered PI was seen for the spleen, gut and brain.

The combination of increased flow velocity and reduced PI in the umbilical arteries supplying the placenta with blood from the fetus, and increased UV return to the fetus implies augmented placental perfusion in the macrosomic fetuses during the second half of pregnancy. This is in keeping with the concept that in these fetuses less of the cardiac output is recycled within the fetal body and more is directed to the placenta31, and in contrast to fetal growth restriction where more of the combined cardiac output is recycled1.

In conclusion, our study shows that non-diabetic macrosomic growth is associated with augmented hemodynamics, particularly on the venous side, with a maintained increase in flow until term, while the arterial side mainly reflects the fact that auto-regulatory systems are not activated. In addition to supporting the concept that UV perfusion of the liver is a growth-driving mechanism in its own right, we believe that our data will be useful for future studies of abnormal excessive fetal growth such as in diabetic pregnancies.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES

C.E. was supported financially by the Western Norway Regional Health Authority (Post doctoral grant no 911581).

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. REFERENCES