Correspondence: Assistant Professor A. Mori, Tokai University School of Medicine, Bohseidai, Isehara, Kanagawa 259–1193, Japan.
Objective To study the relation between diameter pulse (pressure pulse) and flow velocity waveforms in the fetal descending aorta in fetuses with intrauterine growth retardation and acceleration.
Design Blood flow through a tubular system can be expressed by the ratio of blood pressure to vascular resistance. Doppler ultrasound and a phase locked loop echo tracking system coupled to a B-mode ultrasonic imager (central frequency 3.5 MHz) were used to assess downstream resistance and change in blood pressure, respectively.
Setting Tertiary referral unit in a teaching hospital
Participants Serial study between 21 and 40 weeks of 22 women with normally grown fetuses at intervals of four weeks; 25 women with small for gestational age fetuses with uteroplacental insufficiency (high umbilical artery pulsatility index); and six women with large for gestational age fetuses.
Main outcome measures We measured the maximum systolic and minimum diastolic diameter of the fetal descending aorta (the pulse amplitude) and then calculated the pulsatile waveform time integral above the least diastolic diameter (pulsatile area) and total waveform time integral (perfusion area).
Results Normal fetal growth was associated with an increase in systolic and diastolic diameters in the fetal descending aorta with advancing gestational age. Pulse amplitude, pulsatile and perfusion areas increased with gestational age. The increased pulse amplitude and increased pulsatile area in association with a decrease of the pulsatility index in the descending aorta during the second trimester suggested that pulse pressure and stroke volume were increased. In the group with intrauterine growth retardation, systolic and diastolic diameters of the descending aorta and perfusion area were within the normal range. Moreover, there was an increase in the diastolic diameter per unit fetal weight and a decrease in the pulsatile area. In the large for gestational age fetuses, there was an increase in the pulse amplitude and pulsatile area.
Conclusion These findings suggest that in growth restricted fetuses there is an increase in diastolic pressure and a reduction in stroke volume, while in large for gestational age fetuses there is an increase in the pulse pressure and stroke volume. It is possible that measurement of pressure pulse waveform in relation to Doppler velocity waveform may be used to infer changes of blood flow.
Full evaluation of the cardiovascular system involves measurement of the heart rate, peripheral resistance, blood pressure and cardiac contractility. The advent of Doppler ultrasonography has contributed to noninvasive examination of fetal cardiovascular pathophysiology and has been used to study blood flow, especially using the flow-velocity waveform1–3. Arterial velocity waveforms have been studied and interpreted as measurements of downstream resistance. However, examination of blood pressure in the fetus has until now not been possible because of inaccessibility. The pressure pulse waveform propagated along the vascular tree acting laterally on the vessel wall may produce a corresponding change in the diameter of the vessel. It has been reported that, both in animals and humans, the waveform of the diameter of a vessel wall and the pressure propagated along it have an almost identical appearance4,5. Diameter pulse waveforms have therefore been used to infer relative magnitude of blood pressure6. Recording the changing vessel lumen diameter throughout the cardiac cycle provides a method of representing the pressure pulse waveform. Diameter pulse waveforms have been shown to contain abundant haemodynamic information, which can be elicited by monitoring the change in the vessel lumen diameter7. For any isolated length vessel of constant distensibility, the pressure in it is proportional to its volume. Changes in blood pressure will be represented by changes in the diameters of the vessel wall. Blood flow through a tubular system reflects the blood pressure and the resistance.
It is considered that the changing blood flow might influence fetal growth. In order to extend our knowledge of the development of growth disturbances, we investigated the alterations in fetal circulation by this indirect measurement of the arterial pressure pulse, combined with blood flow-velocity waveforms in fetuses with intrauterine growth retardation and large for gestational age fetuses, as compared with fetuses who were appropriate for gestational age.
We routinely studied 22 pregnant women with fetuses who were appropriate for gestational age. In all the women there was ultrasound confirmation of gestational age. Studies were performed between 20 and 40 weeks at 4-weekly intervals. All the women were non-smokers and gave informed consent to participate in the study. One hundred and two sets of measurements were performed in this group. Technically acceptable pulse and flow velocity waveform recordings were obtained in 88 out of 102 occasions, resulting in a rejection rate of 13.7% due to fetal position or movement. All birthweights were appropriate for gestational age (range 2800–3750 g). There were no structural anomalies.
A further 47 pregnant women had fetal growth monitoring in the fetal welfare laboratory because the pregnancy was considered at risk. Of these 16 refused permission to participate in a serial study. The study was performed between 24 and 40 weeks at weekly intervals in the remaining 31 women. One hundred and twenty sets of measurements were performed in this group. The principal associated obstetric complication was maternal hypertension (15), antepartum haemorrhage (2), maternal diabetes mellitus (10) and maternal renal disease (4). In this group 25 fetuses with intrauterine growth retardation and six fetuses who were large for gestational age were selected. Intrauterine growth retardation was defined as:
1no fetal malformation or chromosomal defect.
2ultrasound measurement of fetal abdominal circumference with flattening growth pattern (≤−1.5 SD) below the mean of our reference range.
3pulsatility index (≥−1.5 SD) above the mean of our reference range of Doppler flow-velocity waveform in the fetal descending aorta and umbilical artery, indicating increased downstream resistance.
4postnatal confirmation of a birthweight (≤−1.5 SD) below the mean of Japanese women (range 956–2401 g)8.
The large for gestational age fetuses fulfilled the following criteria:
1no fetal malformation or chromosomal defect.
2ultrasound measurement of fetal abdominal circumference (≥+1.5 SD) above the mean of our reference range.
3postnatal confirmation of a birthweight (≥+1.5 SD) above the mean of Japanese women (range 3875–4519 g)8.
The fetuses with intrauterine growth retardation and those who were large for dates were combined into a growth disturbance group. In the growth disturbance group the gestational age at delivery ranged from 28 to 40 weeks [mean (SD) 34.1 (4.1)]. The last study only was used for analysis and the interval from the last study to delivery was ≤ 7 days [mean (SD) 3.1 (2.1)].
The pressure pulse waveform was recorded from the fetal descending aorta in the lower thorax above the diaphragm. The distance between diametrically opposite points of the vessel lumen was followed using a paired phase locked loop echo tracking system coupled to a B-mode ultrasonic imager. The high sampling frequency (3000 Hz) of our system allowed the waveform to be followed. Each consisted of a gated zero crossing phase detector and a phase-locked loop circuit9. The interval between the points represents the vessel diameter.
The fetus was first imaged in a longitudinal plane and the descending aorta was identified. Tracking markers were then placed on the proximal and distal walls. The electronic markers of each tracking gate were adjusted so as to coincide as close as possible to the inner surface of each wall and then locked in. The distance between the inner vessel walls was measured in a plane perpendicular to the longitudinal axis of the vessel. The ultrasonography signal was displayed on a Macintosh computer. The signal was converted into an analogue output. For analysis of the pulse waveforms the analogue voltage representing the vessel diameter was processed using the computer with a MacPac peripheral.
The following characteristics of the pulse waveform were measured. The first derivative waveform of this waveform was also determined to measure the ventricular ejection time, as previously described10.
1Peak systolic diameter. The maximum aortic diameter recorded during each cardiac cycle (mm).
2End-diastolic diameter. The aortic diameter at the end of the diastolic phase of the pulse waveform (mm).
3Pulse amplitude. The difference between the peak systolic and end-diastolic diameters, expressed in absolute terms (mm).
4Diastolic: systolic diameter ratio. The ratio of the end diastolic to peak systolic diameter. This was expressed as a percentage. This ratio is mathematically related to the amplitude ratio which is (systolic diameter-diastolic diameter)/diastolic diameter
5The cardiac period (pulse duration)
6Pulse waveform time integral was calculated. This area under the pulse waveform is the total area (perfusion area).
7The pulsatile waveform time integral was calculated. This area of the pressure pulse waveform is that which is above the least diastolic diameter (pulsatile area).
8The ventricular ejection time (ms) was calculated using the first derivative waveform
In addition, in all studies, the descending aorta and umbilical artery flow velocity waveforms were recorded as previously described1,3.
The ultrasonic recordings were made during periods of fetal rest without breathing movements. The average of ten consecutive pulse waveforms was calculated for each measurement. For systolic and diastolic diameter measurements, intra- and inter-observer reproducibility was assessed. Real-time images of the fetal descending aorta were displayed in five normal growth fetuses at between 30 and 40 weeks three times for each fetus. The same examiner measured the diameters at intervals of 10 min. Moreover, three observers measured the diameters on each image independently. The intra-observer and inter-observer coefficient of variation was 5.4% and 7.7%, respectively.
The 10th and 90th centile limits of the normal studies were determined using the method of Royston11 and Altman12 which permits a parametric derivation of an age-related variable and allows for a nonlinear relationship between variability and age. Differences for each parameter between the normal and growth disturbance groups were assessed using analysis of variance (ANOVA) with gestational age as a covariable. Fetal weight was estimated from the ultrasound measurement of biparietal diameter, abdominal circumference and femoral length.
All studies were performed with the approval of the Hospital Research and Ethics Committee.
Normal growth group
The pulse duration showed an increase, corresponding to a decreasing mean fetal heart rate from a mean (SD) of 147.2 bpm (6.9) at 21–24 weeks to 130.6 bpm (7.8) at 37–40 weeks. The first derivative of the aortic pulse waveform was used to identify the incisura from closure of the aortic valve. There was no significant change in the mean ventricular ejection time from 0.182 s at 20–23 weeks to 0.184 s at 37–40 weeks. The changing heart rate over this period was associated with a changing diastolic time period. The peak systolic diameter increased linearly from a mean (SD) of 3.56 mm (0.41) at 21–24 weeks to 8.59 mm (0.82) at 37–40 weeks (y = 0.314 χ−3.348, r2= 0.856). The diastolic diameter increased linearly from a mean (SD) of 3.09 mm (0.32) to 7.49 mm (0.76) (y = 0.275 χ−2.96, r2= 0.841). The pulse amplitude also increased in absolute terms from a mean (SD) of 0.51 mm (0.09) to 0.92 mm (0.14) (y = 0.027 χ−0.078, r2= 0.856). The diastolic to systolic diameter ratio showed a very small increase from a mean (SD) of 85.9% (1.2) at to 87.7% (1.5) (y = 0.0182 χ 86.3902, r2= 0.28). The pulsatile area increased linearly from a mean (SD) of 0.09 mm × sec (0.014) at 21–24 weeks to 0.16 mm × sec (0.02) at 37–40 weeks (y = 0.006 χ−0.022, r2= 0.752). The perfusion areas also increased linearly from a mean (SD) of 1.29 mm × sec (0.29) at 21–24 weeks to 2.85 mm × sec (0.48) at 37–40 weeks (y = 0.119 χ−1.491, r2= 0.819). The pulsatility index in the descending aorta and umbilical artery decreased during the second trimester. There was a discrepancy between the pulse amplitude and the pulsatility index of the descending aorta and umbilical artery during the second trimester. Figure 1 shows the developmental changes in the pulse waveforms and velocity waveforms in the descending aorta from 20 to 39 weeks of gestation. The pressure pulse waveforms and blood flow velocity waveforms showed almost identical shapes with advancing gestation, with increasing diastolic flow velocity, increasing perfusion area, increasing diastolic diameter and lengthening of the time in diastole.
Growth disturbance groups
The results from 31 fetuses in which growth disturbance was present were compared with the data from the group with normal growth. Differences for each parameter of the pulse waveform between the normal and growth disturbance groups were examined using ANOVA with gestational age as covariable (Table 1). This is also displayed in the figures using data from the normal group to represent 10th and 90th centile limits. There were no significant changes in mean fetal heart rate and ventricular ejection time between normal and growth disturbance groups (Fig. 2).
Table 1. A comparison between the aortic pressure pulse waveform parameters in normal and growth disturbance groups. The P values shown were calculated by post hoc analysis (Bonferroni test) after normal and growth disturbance groups were compared using ANOVA. EFW = estimated fetal weight.
Small for gestational age (growth restricted) fetuses
Large for gestational age fetuses
Aortic pulse waveform parameter
Systolic diameter (mm)
Systolic diamet er/√EFW(mm/√kg)
Diastolic diameter (mm)
Diastolic/systolic diameter ratio (%)
Pulsatile area (mm × s)
Perfusion area (mm × s)
In the fetuses with intrauterine growth retardation, peak systolic and end-diastolic diameters were within the normal range. The end-diastolic to peak systolic diameter ratio was increased [mean (SD) 91.44% (2.04)](Fig. 3). All values in this group were above the 90th centile value of the normal group. The amplitude of the pulse was decreased relative to the diastolic diameter. It was considered that fetal size might influence the aortic size, as previously described10. The systolic and diastolic diameters of the descending aorta were expressed to the square root of the ultrasound estimate of fetal weight. This was done to allow for the significant differences in size and weight between normal and growth disturbance groups comparable gestational age. Normal ranges were defined. Calculated in this way the diastolic diameter was high in the growth restricted fetuses (Fig. 4), who also had a normal perfusion area (Fig. 5) but a decreased pulsatile area (Fig. 6). An example of a recording of the fetal pressure pulse waveform from a fetus with intrauterine growth retardation at 32 weeks of gestation is shown in Fig. 7 (left panel). In the relationship between the pressure pulse waveforms and blood flow velocity waveforms, the group with intrauterine growth retardation showed a sharper decrease in the velocity waveform during the diastolic phase compared with the pressure pulse waveform. This resulted in low blood flow velocity in diastole, with therefore a high value of the pulsatility index. The slower decline in the pressure pulse waveform before the next heart beat can be seen in Fig. 8 (upper panel).
In the large for gestational age fetuses, peak systolic and end-diastolic diameters were within the normal range. The end-diastolic to peak systolic diameter ratio was decreased [mean (SD) 83.11% (1.33)](Fig. 3). The diastolic diameter to the square root of ultrasound estimate of fetal weight was low in this group (Fig. 4). These fetuses also showed normal perfusion area (Fig. 5) but an increase in the pulsatile area (Fig. 6) with an increase in the pulse amplitude. An example of a recording of the fetal pressure pulse waveform from a large for gestational age fetus at 36 weeks of gestation is shown in Fig. 7 (right panel). In the relationship between the pressure pulse waveforms and blood flow velocity waveforms, the waveforms are almost identical. There was increased systolic flow velocity due to increased pulse amplitude (Fig. 8, lower panel).
We have demonstrated physiological alterations of the fetal circulation during fetal development by comparing arterial pressure pulse waveforms and blood flow velocity waveforms. Blood flow is measured by multiplying the full vessel lumen velocity and the vessel area. Recently, it has been possible to measure mean vessel diameter from monitoring the change in the vessel lumen diameter. However, the velocity profile across the aorta is complex and changes over the cardiac cycle. The maximum velocity will not necessarily be at the centre of the vessel. The full vessel lumen mean velocity may be associated with the Doppler velocity recorded from a sample volume around the centre of the vessel. Accordingly, it is very difficult to measure the full vessel lumen mean velocity.
The pressure and resistance of any fluid travelling through a tubular system can be expressed by the Poiseuille equation13,14:
where ΔP is pressure gradient, r is the vessel radius, η is the viscosity of the fluid and 1 is the vessel length. Resistance is a function of the vessel area and viscosity.
However, it must be recognised that blood is a non-Newtonian fluid and may not obey Poiseuille's law, especially in the capillary and small vessel circulation where there may exist a variety of shear rates. In large arteries flow is far less influenced by viscosity than in small vessels. The inertial forces moving a column of blood during pulsatile flow are more important in large arteries. In small vessels, with parabolic flow profiles, shear between layers is more important. The arteries which we studied have a diameter of more than 3 mm and are therefore considered to be large arteries. Viscosity could be expected theoretically to be less important in determining flow.
Therefore, the Poiseuille equation provides a useful expression of the relation between pressure and resistance. We used Doppler ultrasound and a phase locked loop echo tracking system coupled to a B-mode ultrasonic imager to assess the changing downstream resistance and the changing blood pressure, respectively. A phase locked echo tracking system to record the pulsatile diameter waveform has been used to represent the pulse pressure waveforms.
The principle of an echo-tracking method used for inner vessel diameter changes was based on a phase shift analysis of the zero crossing phase detector of the reflected wave of ultrasound from the vessel wall. Using a 3.5 MHz system at a velocity of sound of 1560 m/sec it can be calculated that each sound wave occupies 0.045 cm. If the movement between the sample intervals is not greater than one half of 0.045 cm, the point whose displacement is being tracked will remain in phase lock. The high-frequency sampling is necessary to ensure that the point displacement between successive samples is not greater than this interval. The major advantage of this system is its high resolution, which far exceeds that of M-mode ultrasonography.
Combining the echo tracking system with B-mode ultrasound imaging makes it possible to study deep lying vessels. We have not measured pressure but rather the pulse pressure waveform. Pulse may be defined as fluctuation in pressure, flow or diameter caused by cardiac contraction. The configurations of the pressure and diameter waveforms are consistent4,5. The magnitude of the pulse pressure will be related with the pulse amplitude in the diameter pulse waveform. The diameter waveform is used to represent the pulse pressure waveform over the cardiac cycle.
In the group with normal fetal growth, the diameter of the descending aorta increases linearly with advancing gestational age. The diameter of the descending aorta is related fairly closely to body surface area. From the pulse waveform in the descending aorta the pulse amplitude and pulsatile area increase linearly with advancing gestational age. The pulsatility index of the descending aorta decreases during the second trimester. These results might reflect increasing stroke volume by increasing pulse pressure and decreasing systemic vascular resistance produced by growth of the placenta.
In the growth disturbance groups, the aortic pressure pulse waveform showed some consistent differences. Although absolute diastolic diameter did not differ between the normal and growth disturbance groups, the diameter per unit fetal weight was increased in the group with intrauterine growth retardation due to increased systemic vascular resistance caused by placental insufficiency. The end-diastolic to peak systolic diameter ratio was increased. The amplitude of the pulse was decreased relative to the diastolic diameter. The pulsatile area which is above the least diastolic diameter was decreased in the growth restricted fetuses. Moreover, the perfusion area (total area) were within normal range due to increased diastolic diameter. Although increased blood pressure is accompanied by an increased diameter7 and pulsatile area is associated with stroke volume15, we suggest that with intrauterine growth retardation there is an increase in diastolic pressure and a reduction in stroke volume resulting in decreased pulse pressure. On the other hand, the diastolic diameter per unit fetal weight was decreased in the large for gestational age group suggesting that the systemic vascular resistance decreased. The diastolic to peak systolic diameter ratio was decreased. The amplitude of the pulse was increased relative to the diastolic diameter. Although the pulsatile area was increased with the increased pulse amplitude, the perfusion area was within the normal range due to the decreased diastolic diameter. Since increased aortic blood flow is reflected by an increased pulse pressure (pulse amplitude)15, we suggest that there is an increase in stroke volume resulting in the increased pulse pressure in the large for gestational age fetuses. With intrauterine growth retardation therefore there is a reduction in the pressure gradient in the descending aorta and an increase in downstream resistance, which from Poiseuille's equation, results in decreased blood flow in the descending aorta.
With large for gestational age fetuses the increased pressure gradient and the decreased systemic vascular resistance in the descending aorta by itself, according to Poiseuille's equation, will result in increased blood flow in the descending aorta. In our study the number of fetuses with growth acceleration was a small. The inference in the changes of blood flow could be more apparent with a larger number of fetuses with growth disturbance, as well as fetal animal models.
The contour of the arterial pressure waveform has been the subject of much study in the cardiovascular literature. It has been explained by the concept of standing waves. The effects of changes in pulse wave reflection on the configurations of pressure and flow have been studied. The similarity between the shapes of aortic pressure waveforms and the velocity waveforms increases when blood flow increases due to a reduction of peripheral resistance and an increase of arterial pulse pressure6. We observed that the similarity between the shapes of pressure pulse waveforms and the flow velocity waveforms in normal growth group increased with advancing gestational age. These waveform shapes showed an almost identical appearance at 39 weeks of gestation (Fig. 1).
The perfusion area of the growth disturbance group was not different from the normal growth group. The measured pressure pulse and flow velocity waveforms consist of forward and reflected waves. The reflection of the systolic wave from the periphery occurs to produce the diastolic wave, according to the concept of a standing wave. The systolic part of the pressure wave ends at the incisura which is caused by cardiac relaxation at the end of systole. The wave during diastole is seen superimposed on the general decline in pressure before the next heart beat due to the result of reflection16. The pressure pulse waveform of fetuses with intrauterine growth retardation showed this waveform shape and a decrease in the pulsatile area. However, the perfusion area in this group was within the normal range due to an increased diastolic diameter. If peripheral resistance is high then the wave is strongly reflected and returns to cancel the latter part of the blood flow velocity waveform. This results in zero, and therefore low blood flow velocity in diastole. The configurations of both waveforms in growth retardation showed a dissociation which is a relatively sharper decrease in the blood flow velocity waveform than the pressure pulse waveform in diastole. In contrast, the pressure pulse waveform of the large for gestational age group showed a sharper increase in the early systolic phase and an increase in the pulsatile area with an increase in the pulse amplitude. The perfusion area in this group was within the normal range due to a decreased diastolic diameter. A low resistance bed allows onwards travel of the pressure pulse waveform and thus high pulse amplitude resulting from the increased pulse pressure16,17. The large for gestational age group showed the increased systolic blood flow velocity due to a high pulse pressure. It is considered that the most efficient circulation, with the least reflection, is obtained when those waveform shapes resemble each other. We observed the close resemblance of those waveform shapes in this group from gestational week 35. Animal data is necessary to extend our understanding of the relationship between shapes representing the pressure pulse and flow velocity waveforms.
We conclude that ultrasonic recordings of the changes in arterial diameter during the cardiac cycle yield important information about the fetal circulation. It is possible that measurement of pressure pulse waveforms in relation to the Doppler velocity waveform may be used to infer changes of blood flow. The most efficient circulation, with the least reflection, is characterised by the similarity between the pressure pulse and flow velocity waveform shapes.