Longitudinal quantification of uterine artery blood volume flow changes during gestation in pregnancies complicated by intrauterine growth restriction


* Dr J. C. Konje, Fetal Growth and Development Research Group, Department of Obstetrics and Gynaecology, Leicester Royal Infirmary, Robert Kilpatrick Clinical Sciences Building, Leicester LE2 7LX, UK.


Objective To quantify and compare longitudinal uterine artery volume flow changes in appropriate for gestational age (AGA) pregnancies and those complicated by intrauterine growth restriction (IUGR).

Design Serial longitudinal study.

Setting Large UK Teaching Hospital Obstetrics and Gynaecology Department and Institute of Anatomy, RWTH Aachen, Germany.

Population Pregnant women with accurately dated singleton pregnancies.

Methods Quantified uterine volume flow was prospectively measured by colour power angiography in (a) 32 women with abnormal uterine artery Doppler velocimetry at 20 and 24 weeks of gestation and with risk factors for IUGR (IUGR group) and (b) 25 women with normal uterine artery Doppler velocimetry and no risk factors for IUGR (AGA group) between 20 and 38 weeks of gestation. Values obtained from each gestation were compared using unpaired t test.

Main outcome measures Gestational age at delivery, birthweight and total quantified volume flow (mL/min) per gestation in IUGR and AGA pregnancies.

Results Twenty of the 32 women recruited into the IUGR group and 18 of the 25 controls fulfilled the criteria for inclusion in the analyses. The mean birthweight [SD] and gestational age [SD] at delivery in the IUGR and AGA groups were 2634 [277] versus 3429 [349] g and 39.5 [1.2] and 41.1 [2.3] weeks, respectively. The diameter of the proximal uterine artery just after it crosses the external iliac artery was smaller in the IUGR group from as early as 20 weeks of gestation but this difference only became statistically significant from 24 weeks of gestation and widened as pregnancy advanced. The quantified volume flow in the IUGR group was significantly less than that in the AGA throughout the study period (287 [117] versus 328 [159] mL/min at 20 weeks (P < 0.05), 334 [169] versus 538 [181] mL/min at 24 weeks [P < 0.004] and 534 [332] versus 830 [284] mL/min at 38 weeks of gestation [P < 0.002]). Volume flow in the IUGR group was 12.5% and 36.7% less than that in the AGA group at 20 and 28 weeks of gestation, respectively.

Conclusion Proximal uterine artery diameter and quantified volume flow change with gestation and show significant differences between AGA pregnancies and those complicated by IUGR. These changes occur early and become more marked as pregnancy advances. Early use of these measurements may identify pregnancies at risk of complications.


During pregnancy, the uteroplacental circulation increases dramatically to meet the metabolic demands of the uterus and its contents. A poor or inadequate circulation is recognised to be associated with fetoplacental compromise1. Assessing this circulation may therefore offer a useful means of fetomaternal surveillance in at risk pregnancies. Direct measurements of uterine artery blood volume flow have been limited by lack of reliable non-invasive techniques and have therefore been measured indirectly by means of Doppler velocimetry.

Significant improvements in Doppler ultrasonography techniques have enabled volumetric quantification of flow in the uterine circulation in vivo transvaginally2,3 or transabdominally4, but these techniques have significant limitations. Conventional colour Doppler, for example, does not delineate vessel diameter accurately and blooming may affect the accuracy of measurement where the colour extends beyond the vessel wall. Recently, volume flow in small fetal arteries5 and maternal uterine arteries in uncomplicated pregnancies6 has been quantified non-invasively using colour power angiography. This technique not only enables easy identification of the vessels but allows effective insonation without the effects of vessel wall movement on the Doppler signal acquired5,7. We recognised that there may be some concerns about the timing of the measurements in relation to the cardiac cycle (e.g. during systole or diastole) on the vessel diameter especially that a 10% difference in diameter could affect the estimated volume flow by as much as 25%8. However, since values we previously obtained from normal pregnancies6 were very close to those obtained by electromagnetic flowmetry at hysterotomy9 and as we consistently made the measurements during systole, the influence of the cardiac cycle on the measurements obtained must be considered to be minimal.

In this serial longitudinal study, we used colour power angiography to quantify changes in the uterine artery volume flow in pregnancies complicated by intrauterine growth restriction (IUGR) and then compared them to values obtained from appropriate for gestational age (AGA) pregnancies. The measurements were made from a well-defined reference point4 and therefore by-passed the problem of variability in size along the course of the vessel.


Women with singleton pregnancies at risk of IUGR were recruited after the 24 week ultrasound scan. These were all patients attending the Fetal Growth Clinic with various risk factors for IUGR. The risk factors were mainly previous IUGR fetuses, low maternal weight and recurrent miscarriages (not due to thrombophilia). The pregnancies had been accurately dated by last menstrual period and crown rump length measurements between 8 and 12 weeks of gestation. If there was a discrepancy of more than 10 days between the ultrasound estimated gestation and the last menstrual period gestational age, the pregnancy was dated from the ultrasound scan measurements.

The Fetal Growth Clinic protocol requires all patients being scanned at 20 weeks to have uterine artery Doppler velocimetry. Pregnancies with abnormal uterine artery Doppler flow waveforms or indices are automatically rescanned at 24 weeks for fetal biometry, amniotic fluid index estimation and repeat uterine artery Doppler velocimetry. Those pregnancies with persisting abnormal uterine artery Doppler indices as defined by Bower et al.4 are followed up serially until delivery. The follow up studies are performed at 28 weeks and thereafter every two weeks until delivery. During each insonation including that at 20 weeks of gestation, the uterine artery volume flow was quantified as described below. Informed consent for participation in the study was obtained from each patient.

We set out to determine the volume flow in the IUGR pregnancies only (having previously determined values in AGA pregnancies). We, therefore, retrospectively plotted the growth patterns of all the pregnancies, and only those that exhibited crossing of centile lines in their growth were included in the IUGR group. In addition to this, birthweight had to be below the 10th centile for gestational age and gender for the Midlands, UK10. Pregnancies complicated by severe IUGR (defined as birthweight below the third centile for gestational age, gender and race) and an amniotic fluid index below the third centile were excluded as they are invariably delivered before 38 weeks of gestation. All pregnancies ending before 36 completed weeks of gestation or in which both uterine arteries could not be measured or had less than four measurements were also excluded. It was necessary to restrict the inclusion (in this preliminary study) to mild IUGRs to allow measurements in the late third trimester to be made as most cases of severe IUGR tend to be delivered early. In addition, the heterogeneous nature of severe IUGR makes the identification of similar phenotypes difficult.

Twenty-five women with no risk factors for IUGR or hypertensive disorders of pregnancy and who had normal uterine artery Doppler at 20 weeks of gestation were recruited as controls (AGA group). They had serial ultrasound scans at similar intervals to those in the study group. For the purposes of the comparative analysis, only those whose fetuses exhibited a normal growth trajectory, were delivered after 38 weeks of gestation and had at least four Doppler measurements were included. None of the subjects in this study had been included in our previous study6 of normal subjects.

To quantify volume flow through the uterine arteries, the vessel was identified and the Doppler waveforms acquired as previously described4,6. The cross sectional area of the proximal uterine artery was calculated from a measurement of its internal diameter just after it crosses the external iliac artery. In-built software then calculated the time-averaged mean velocity from three to five completed cardiac cycles, and by multiplying this by the area, the volume flow in millilitres per minute was obtained. All the Doppler studies were undertaken by one operator to minimise variability.

All the measurements were made with an Advanced Technology Laboratory (ATL) High Definition Imaging (HDI) 3000 ultrasound machine (ATL, Bothwell, Washington, USA) using a 4.3 mHz curvilinear probe. The settings were as previously described6. Measurements were made with the patient in the semirecumbent position and slightly tilted to the left. The power setting was at its lowest output (<92 mW/cm2 spatial peak temporal average intensity). The wall filter was set low at 75 mHz so that all diastolic velocities could be detected. The pulse repetitive frequency was set at 700 Hz, the colour gain at 100%, colour sensitivity at 16 and the persistence at medium. We have previously shown that these settings provide the advantage of marked improvements in the ease with which the vessels are identified6. The intra-observer coefficient of variation was similar to that previously reported in our paper on normal pregnancies (5–8% for the diameter of the uterine arteries and 3–5% for the volumetric measurements). Each uterine artery was measured twice and the average of the two measurements was taken for the volume flow on each side. The total volume flow to the uterus was the sum of the mean from the measurements on each side. Fetal biometric indices were used to estimate fetal weight11 at each gestation and quantified volume flow per kilogram estimated fetal weight was determined.

Results are presented as means and one standard deviation of the total uterine artery volume flow (right and left) to the uterus for each gestational time point. Measurements at the different gestational ages in the AGA and IUGR groups were compared using an unpaired t test at the 5% level of significance.


Thirty-two women were recruited into the study group. Twelve were excluded after delivery since they did not meet the inclusion criteria—seven were delivered preterm because of severe pre-eclampsia (3), prelabour rupture of fetal membranes (1) and very severe IUGR requiring delivery between 29 and 30 weeks (2) and at 32 weeks (1). The remaining five were full term AGA pregnancies and were therefore excluded. All those with severe pre-eclampsia were complicated by IUGR. Of the 25 controls recruited, only 18 satisfied the inclusion criteria. Seven were excluded because of the inability to study both uterine arteries (2), delivery before 38 weeks of gestation (2) and fewer than three measurements (3).

The demographic characteristics of the 20 IUGR pregnancies and the 18 controls are shown in Table 1. As expected, the mean birthweight at delivery was statistically greater in the control group (3429 [349] versus 2643 [277] g, P < 0.05), but the mean gestational age was not significantly different although it was slighter longer in the controls. At the 24th week scan, there were no significant differences in the biometric indices in the two groups although the abdominal circumference was smaller in 25% of cases in the IUGR group.

Table 1.  Demographic characteristics of the IUGR and AGA groups. Values are given as means [SD], n (%) and ratio.
Demographic variableIUGR group (n= 20)AGA group (n= 18)
  1. RI = resistance index.

  2. *P < 0.05.

Maternal age (years)26.4 [4.5]27.1 [5.2]
Gestational age at delivery (weeks)39.5 [1.2]41.1 [2.3]
Birthweight (g)2643 [277]3429 [349]*
RI at 24 weeks of gestation0.83 [0.21]0.58 [0.18]*
Smokers9 (45)6 (37.5)
Male/female ratio9:119:7

The diameter of the proximal uterine arteries and the volume flow through them are shown in Tables 2 and 3, respectively. Differences in the diameter of the uterine arteries (Table 2) and the quantified volume flow (Table 3) were present as early as 20 weeks although they were not statistically different at that gestation. A statistically significant difference was, however, observed at 24 weeks and persisted until the end of the pregnancy. At 20 weeks, the volume flow in the IUGR group was 12.5% less than that in the AGA group but this difference more than doubled to 36.7% at 24 weeks and remained around this level until delivery. Quantified volume flow in both the IUGR and AGA groups increased throughout gestation. The increase in the IUGR group was less marked throughout the study period compared with that in the AGA group. The quantified volume flow per arterial cross section (i.e. per mm2), which is depicted in bold in Table 3, was similar in both groups at all stages except at 32 weeks of gestation. There was a steep fall from 20 to between 30 and 32 weeks of gestation after which it remained stable until 38 weeks of gestation. Quantified volume flow per estimated kilogram of fetal weight decreased steeply until term (Table 4). There were no statistically significant differences in quantified volume flow per estimated fetal weight in the two groups throughout gestation.

Table 2.  Changes in proximal uterine artery diameter (mm) with gestation in IUGR and AGA groups. Values are given as means [SD].
Gestational ageIUGR group (n= 20)AGA group (n= 18)
  1. *P < 0.05.

202.3 [0.2]2.5 [0.1]
243.0 [0.8]3.6 [0.6]*
283.4 [0.6]4.3 [0.8]*
303.7 [0.8]4.7 [0.7]*
324.3 [0.9]5.0 [0.8]*
344.2 [0.7]5.1 [0.6]*
364.3 [0.6]5.3 [0.6]*
384.3 [0.8]5.4 [0.7]*
Table 3.  Changes in uterine artery quantified volume flow (mL/min) with gestation in IUGR and AGA pregnancies. Values are given as means [SD]. Quantified volume flow (mL/mm2) per cross sectional area of the proximal uterine artery is shown in bold.
Gestation (weeks)IUGR Group (n= 20)AGA group (n= 18)
  1. *P < 0.05.

20287 [117] 69.1328 [159] 55.8
24334 [169] 47.5528 [181]*51.9
28357 [217] 39.3600 [211]*41.3
30413 [213] 38.4645 [199]*37.2
32435 [198] 29.8711 [211]*36.2*
34512 [223] 36.9769 [233]*37.6
36535 [245] 38.6801 [265]*36.3
38534 [332] 38.5830 [284]*36.2
Table 4.  Quantified volume flow (mL/kg/min) of estimated fetal weight. Values are presented as mean [SD]. There were no statistical differences in the two groups throughout the study period.
Gestation (weeks)IUGR group (n= 20)AGA group (n= 18)
201289 [524]1384 [513]
24894 [301]907 [288]
28601 [228]624 [210]
30545 [214]561 [192]
32412 [132]432 [114]
34326 [87]358 [98]
36301 [69]309 [79]
38294 [67]302 [48]


Trophoblastic invasion of the distal uterine and spiral arteries with concomitant dilatation of the lumens results in an increase in uteroplacental circulation12. Failure or defective endovascular trophoblast invasion results in a reduction in the uteroplacental circulation and has been suggested to be involved in impaired fetal growth13. These changes, however, are restricted to the endometrial and innermost myometrial segments of the spiral arteries. By contrast, the extramyometrial (proximal) segments of the uterine arteries, where our measurements were taken, are not invaded by trophoblast and show less impressive changes in luminal width. Since uterine artery Doppler velocimetry indirectly assesses impedance of flow through the vessels, high resistance indices in the proximal segments may indicate reduced luminal width of the respective segment and/or inadequate adaptation of its downstream segments (spiral arteries and intervillous space).

In our highly selected IUGR cohort (32 cases) at recruitment with risk factors for IUGR and abnormal uterine artery Doppler velocimetry, only a proportion of the pregnancies (81.3%) were indeed growth restricted. Twelve and a half percent were complicated by pre-eclampsia. Previous studies in low risk populations have shown that while abnormal uterine artery Dopplers were sensitive in identifying the high risk population, the positive predictive value for either IUGR or IUGR and pre-eclampsia was low14,15. A much higher incidence of IUGR in our study group could, however, be ascribed to the inclusion criteria of abnormal uterine artery Dopplers at 24 weeks of gestation associated with risk factors for IUGR.

Although the proximal arterial diameters in the IUGR group were smaller than those in the AGA group, these differences at 20 weeks of gestation were not statistically significant. Indeed, the differences in diameter between the IUGR and AGA groups became significant only at 24 weeks of gestation. A closer examination of changes in the diameter of the uterine arteries and the quantified volume flow through them show an interesting pattern. It appears that in the IUGR group, changes in flow and diameter occurring after 20 weeks of gestation were steady but much more modest compared with the AGA group.

Interestingly, in all stages studied, the flow per square millimetre of proximal arterial cross section was the same in both groups: it decreased from the 20th to the 28th week, then reached an identical plateau in both groups from the 30th to the 38th week. At first glance, this finding is surprising since one might expect lower flows per square unit in IUGR pregnancies as the downstream parts of these arteries are mal-dilated and show increased impedance. To the best of our knowledge, the control of the ‘well-known pregnancy-induced arterial dilation upstream’ by trophoblast invasion of the spiral arteries has not yet been sufficiently explained. In guinea pig experiments, Moll et al.16 showed that prior to trophoblast invasion, oestrogens might have already stimulated vessel growth and dilatation. However, the more likely explanation for our findings of comparable flows per square unit proximal arterial cross section in AGA and IUGR is that arterial width of the upstream segments of the uterine arteries only secondarily adapts to the flow requirements defined by trophoblast invasion and dilatation of the downstream segments (spiral arteries).

What may be deduced from these observations? The continuous increase in the diameter of the uterine arteries in both groups suggests that the arterial adaptation to pregnancy is not only in the spiral arteries but also in the proximal segments of the uterine arteries and continues throughout gestation. In the IUGR group, however, this process appears to be either impaired or slowed down considerably (as early as the second trimester). This could be responsible for the significantly lesser increase in diameter and volume flow. It is still an open question (a) whether reduced arterial dilatations of the proximal uterine artery (induced by humoral factors) and of the more distal spiral artery (related to endovascular trophoblast invasion) are independently controlled processes that both contribute to reduced uteroplacental blood flow in IUGR or (b) whether the missing upstream dilatation is simply an adaptation to reduced flow in this vessel caused by increased downstream impedance.

On the bases of our results, we speculate that there is a complex interplay among the fetus, the placenta and the haemodynamic changes in the uteroplacental circulation in complicated pregnancies. Impedance in the uteroplacental circulation very likely does not exclusively depend on the luminal width of the uteroplacental arteries but also on the width of the intervillous space and consequently, the numerical density of placental villi is an important factor. Placentas from IUGR show reproducible aberrations from normal villous development17,18. Villous density is strongly reduced and intervillous space enlarged in IUGR with absent or reverse end diastolic flow in the umbilical arteries, but it is steeply increased as compared with normal in cases with preserved end diastolic flow. In both groups, the placentas and the villous trees are smaller, and consequently, the intervillous length to be passed by the blood is smaller as compared with normal conditions. However, both clinical entities show reduced uteroplacental flows. According to Kingdom and Kaufmann19 and Khaliq et al.20, fetoplacental angiogenesis and, therefore, villous developmental patterns reflect intraplacental oxygenation. The latter is also considered to be an important regulator of trophoblast invasion21.


In conclusion, we have demonstrated that in IUGR pregnancies, not only are uterine artery Doppler indices different but total quantified volume flow is significantly less than that in AGA pregnancies. In addition, in IUGR pregnancies, not only does the distal segment of the uteroplacental arteries (spiral arteries) show mal-adaptation to pregnancy but the proximal upstream segments also exhibit less dilatation. Measuring flow in these segments, therefore, reflects exactly the events taking place in the spiral arteries. The volume flow per estimated kilogram of fetal weight remains the same irrespective of the fetal growth pattern, suggesting that the fetoplacental unit (intervillous space, placental villi and their fetoplacental vascular system) contribute to the modulation of uteroplacental circulation rather than the latter modulating fetal growth exclusively. We are inclined to conclude from these results that volume flow may be a better means of identifying pregnancies at high risk for IUGR. Although this study concentrated mainly on pregnancies with mild IUGR, we hope to investigate those complicated by severe IUGR in future.

In the application of colour power angiography to volumetric studies, consideration must be given to the potential pitfalls of this technique. We have highlighted some of these in the Introduction. While measurements from this and previous studies have suggested that it is as accurate as standard thermodilution techniques, we would like to see our values duplicated by others. It is only after such confirmatory studies that the relevance of these four findings to clinical practice could best be assessed.