SEARCH

SEARCH BY CITATION

Keywords:

  • Doppler;
  • uterine artery;
  • uteroplacental circulation;
  • volume blood flow

Abstract

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

Objective

To test the hypothesis that Doppler-derived (calculated) uterine artery volume blood flow (cQUtA) reflects accurately volume blood flow measured directly (mQUtA) in an experimental setting.

Methods

Five pregnant sheep were instrumented at 122–130 days of gestation under general anesthesia. After a 4-day recovery period, maternal hemodynamics were varied by administering to the sheep under general anesthesia noradrenaline, beta-blocker, low oxygen gas mixture, epidural bupivacaine and ephedrine, consecutively. The central venous pressure was obtained with the help of a thermodilution catheter. The mean arterial pressure and acid–base status were monitored using a 16-gauge polyurethane catheter inserted into the descending aorta via a femoral artery. A 6-mm transit-time ultrasonic perivascular flow probe was used to measure the mQUtA. Doppler ultrasonography of the uterine artery was performed and volume blood flow was obtained simultaneously by the transit-time ultrasonic perivascular flow probe during each phase of the experiment.

Results

A total of 31 observations were made. The mQUtA varied between 90 and 800 (mean ± SD, 419 ± 206) mL/min during the experiments. The corresponding values for the cQUtA were 110 and 900 (mean ± SD, 459 ± 211) mL/min. There was a significant correlation (R = 0.76; P < 0.0001) between mQUtA and cQUtA. The mQUtA correlated positively with Doppler-derived uterine artery absolute velocities, i.e. peak systolic (R = 0.50; P = 0.004), end-diastolic (R = 0.53; P = 0.002) and time-averaged maximum (R = 0.69; P < 0.0001) and time-averaged intensity weighted mean (R = 0.75; P < 0.0001) velocities.

Conclusion

cQUtA correlates well with volume blood flow measured directly. Doppler-derived uterine artery absolute blood flow velocities reflect uteroplacental volume blood flow in pregnant sheep. Copyright © 2007 ISUOG. Published by John Wiley & Sons, Ltd.


Introduction

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

Impaired placental perfusion due to inadequate trophoblastic invasion of the spiral arteries is associated with pre-eclampsia and intrauterine fetal growth restriction1. Increased pulsatility or resistance indices and/or presence of protodiastolic notching in the uterine artery Doppler velocity waveforms are used in clinical practice to identify pregnancies at risk of developing such complications. However, if the technical difficulties could be overcome, it would be more appropriate to assess perfusion and vascular resistance of the maternal side of the placenta by measuring uteroplacental volume blood flow.

The volume blood flow in a vessel can be calculated if the spatial mean velocity and cross-sectional area (CSA) of the vessel are known. However, non-invasive measurement of volume blood flow has its limitations2, 3, related to angle dependency of Doppler-derived blood velocity waveforms, error in the diameter measurement of the vessel, and the limited capability of ultrasound equipment to estimate accurately the spatial mean velocity. Several reports have been published on the possibility of uterine artery volume blood flow (QUtA) measurement4–10. Recent developments in ultrasound technology, such as color power angiography, provide new possibilities for refining this technique5. However, non-invasive measurement of QUtA has not been validated experimentally.

We tested the hypothesis that Doppler-derived (calculated) uterine artery volume blood flow (cQUtA) reflects accurately the instantaneous volume blood flow measured directly using a perivascular flow probe (mQUtA) in an experimental setting under varying volume blood flow conditions.

Materials and Methods

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

Measurements were made on five pregnant sheep (gestational age, 122–130 days; term, 145 days) that were part of a research protocol investigating the effect of antihypertensive therapy on maternal and fetal circulation. All experiments were performed in accordance with the guidelines of the European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes (1986) and in compliance with the European Union Directive 86/609/EEC (1997). The research protocol was approved by the Animal Care and Use Committee of the University of Oulu, Finland.

Surgical procedure and instrumentation

Before surgery, food was withheld for 18 h. Premedication consisted of intramuscular ketamine (2 mg/kg) and midazolam (0.2 mg/kg). General anesthesia was induced with intravenous propofol (4–7 mg/kg). The anesthesia was maintained with isoflurane (1–2.5%) in an oxygen–air mixture delivered via an endotracheal tube, combined with intravenous boluses of fentanyl as required. Mechanical ventilation was maintained throughout the surgical procedure with a Siemens 730 ventilator (Siemens-Elema AB, Solna, Sweden). An auricular artery was cannulated to measure the arterial blood pressure (BP) and heart rate. An 8.5-F (2.8-mm) Introflex polyurethane sheath with bonded Touhy-Borst valve and a sideport (Edwards Lifesciences LLC, Irvine, CA, USA) was inserted into the left external jugular vein for intravenous access and for the insertion of a thermodilution catheter on the experimental day.

A midline laparotomy was performed and a 6-mm transit-time ultrasonic flow probe (Transonic Systems Inc., Ithaca, NY, USA) was placed around the uterine artery supplying the pregnant uterine horn proximal to its bifurcation. The flow probe was secured in a fixed position with sutures, tunneled subcutaneously, and exteriorized through a small incision in the ewe's flank. The laparotomy incision was closed. For postoperative analgesia, a fentanyl patch was attached to the ewe's tail, and intramuscular fentanyl (2 µg/kg) was given in the first 48 h and buprenorphine (0.01 mg/kg) thereafter. The ewes were given 1 g ampicillin daily via the jugular venous catheter, which was flushed daily with heparinized saline. The animals were allowed to recover for 4 days before experiments started.

Data acquisition

Data acquisition was carried out on anesthetized animals on the 5th postoperative day. General anesthesia was induced with intravenous propofol (4–7 mg/kg) and maintained throughout the experiments with isoflurane (1–1.5%) in an oxygen–air mixture delivered via an endotracheal tube.

A 16-gauge polyurethane catheter was inserted into the descending aorta via a femoral artery for monitoring the aortic BP and acid–base status. A thermodilution catheter (Criticath SP5107H, Becton Dickinson, Sandy, UT, USA) was inserted into the pulmonary artery through the external jugular vein access. In addition, a 19-gauge epidural catheter was placed into the epidural space just above the lumbosacral junction. Ringer's lactate solution was infused freely until pulmonary capillary wedge pressure (measured by filling the balloon located close to the tip of thermodilution catheter with 1.5 mL air to occlude a pulmonary capillary artery) reached a value of 6 mmHg, and thereafter at a fixed rate of 200 mL/h. The sheep was then positioned supine with a right lateral tilt.

The animals were allowed to stabilize for 30 min before the baseline (Phase 0) measurements were made. In Phase I, noradrenaline infusion was started at an initial rate of 0.2 µg/kg/min and adjusted to achieve a systolic BP of at least 30% above the baseline, or over 140 mmHg. The maternal hemodynamic and uterine artery Doppler measurements were obtained 15 min after achieving the targeted increase in maternal systolic BP. In Phase II, the ewe received an infusion of a beta-blocker (labetalol 1mg/kg) over 30 min in order to decrease the systolic BP to baseline level. In Phase III, hypoxemia was induced by replacing oxygen with medical air (3 L/min) in the rebreathing circuit until the maternal oxyhemoglobin saturation level reached 80–90%. In Phase IV, the ewe was allowed to recover from hypoxemia by returning the inhaled oxygen concentration to baseline. In Phase V, 0.5% bupivacaine was administered through the epidural catheter to a total dose of 0.3 mL/kg body weight 2 min after an initial test dose of 5 mL. Hypotension was allowed to develop to at least a 20% decrease in the maternal systolic BP. In Phase VI, 5-mg boluses of ephedrine were administered intravenously to raise the maternal systolic BP to at least 90% of the baseline level. The total experiment lasted for an average of 3 h.

The aortic BP of the ewe was monitored continuously using a disposable pressure transducer (DT-XX, Ohmeda, Hatfield, UK). The mean arterial pressure (MAP) was calculated as: MAP = diastolic BP + 1/3(systolic BP − diastolic BP). The central venous pressure (CVP) was measured continuously from the superior vena cava just above the right atrium with the thermodilution catheter, which has an extra lumen for CVP measurement approximately 30 cm proximal to the catheter tip. The mQUtA was measured directly with a perivascular transit-time ultrasonic flow probe (T206, Transonic Systems Inc., Ithaca, NY, USA). All these variables were recorded continuously at a sampling rate of 100 Hz using a polygraph (UIM100A, Biopac Systems Inc., Santa Barbara, CA, USA) and computerized data acquisition software (Acqknowledge v. 3.5.7 for Windows, Biopac Systems Inc., Santa Barbara, CA, USA).

Ultrasonography was performed, using a Vivid 7 Dimension ultrasound system (GE Vingmed Ultrasound, Horten, Norway) with a 4-MHz curvilinear probe, during each phase of the experiment. The high-pass filter (low velocity reject) was set at minimum (1 cm/s). Color Doppler was used to identify the uterine artery and to optimize the insonation angle (kept < 20°) for pulsed-wave Doppler interrogation. Blood flow velocity waveforms were obtained from the proximal portion (before bifurcation) of the same uterine artery around which the perivascular flow probe was attached (Figure 1a). Online measurements were performed using the software supplied with the ultrasound machine and the following parameters were obtained: peak systolic velocity (PSV), end-diastolic velocity (EDV), time-averaged maximum velocity (TAMXV), time-averaged intensity-weighted mean velocity (TAMEANV) and heart rate (HR). An average value of three consecutive cardiac cycles was used for statistical analysis.

thumbnail image

Figure 1. (a) Uterine artery blood velocity waveforms obtained by Doppler ultrasonography and (b) measurement of uterine artery diameter in a pregnant sheep using power Doppler angiography.

Download figure to PowerPoint

The diameter of the uterine artery was measured during systole using power Doppler angiography (Figure 1b) as described by Konje et al.5. Measurements were repeated three times and the average value was used for the calculation of uterine artery CSA, assuming the vessel to have a circular lumen: CSA = π(diameter/2)2. cQUtA was calculated according to the following formula: cQUtA(mL/min) = TAMEANV (cm/s) × CSA (cm2) × 60.

During each phase of the experiment, maternal hemodynamic and uterine artery Doppler parameters were obtained simultaneously and maternal arterial blood samples were taken and analyzed for acid–base status (corrected for a body temperature of 39 °C) using an i-Stat® 1 Analyser (i-Stat Corporation, East Windsor, NJ, USA). At the end of the experiment the animals were euthanized using a lethal dose of pentobarbital.

Data were analyzed using Statistical Software for Social Sciences for Windows version 13.0 (SPSS Inc. Chicago, IL, USA). Linear regression analysis was used to examine the association between variables. Correlations were tested using Pearson's correlation coefficient. Agreement between the two methods of QUtA measurement was assessed for bias and precision using Bland–Altman analysis11, 12 on log-transformed data. Bias was defined as the geometric mean of all volume flow measurement ratios (cQUtA/mQUtA) and precision as 95% confidence limits of agreement. The statistical significance level was set at a P-value of < 0.05.

Results

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

A total of 31 observations were made in five ewes. The average weight of the ewes was 62 (range, 50–73) kg. The mQUtA varied between 90 and 800 (mean ± SD, 419 ± 206) mL/min during the experiments. The corresponding values for the cQUtA were 110 and 900 (mean ± SD, 459 ± 211) mL/min. During different phases of the experiment, the mean values of the maternal MAP ranged between 68 and 126 mmHg, CVP ranged between 4.80 and 6.67 mmHg, arterial oxygen saturation between 78 and 97%, and pH between 7.32 and 7.44. The Doppler ultrasonographic parameters of uteroplacental circulation measured non-invasively are presented in Table 1.

Table 1. Uterine artery blood flow parameters measured by Doppler ultrasonography during different phases of the experiment
 Phase
Variable0 Baseline (n = 5)I Noradrenaline (n = 5)II Beta-blocker (n = 5)III Hypoxemia (n = 5)IV Recovery (n = 4)*V Hypotension (n = 4)*VI Ephedrine (n = 3)*
  • Values are mean (SD).

  • *

    Missing values due to technical difficulties during data collection. cQUtA, Doppler-derived uterine artery volume blood flow; EDV, end-diastolic velocity; PSV, peak systolic velocity; TAMEANV, time-averaged intensity-weighted mean velocity; TAMXV time-averaged maximum velocity; UtA, uterine artery.

PSV (cm/s)99.8 (24.7)101.0 (26.8)91.9 (19.7)95.7 (22.5)100.8 (26.4)75.2 (37.2)99.8 (40.8)
EDV (cm/s)48.7 (19.2)51.1 (15.2)28.2 (14.1)43.1 (7.8)46.6 (4.4)20.8 (11.3)60.2 (8.4)
TAMXV (cm/s)67.1 (16.4)68.0 (20.9)55.6 (6.4)61.6 (11.0)66.7 (13.6)41.2 (10.9)82.0 (16.1)
TAMEANV (cm/s)38.9 (6.5)37.5 (14.0)30.6 (6.3)34.9 (9.6)37.8 (13.3)24.5 (7.2)45.6 (7.8)
Heart rate (bpm)121 (23)122 (13)98 (8)98 (10)99 (9)87 (11)99 (17)
UtA diameter (cm)0.53 (0.11)0.53 (0.09)0.55 (0.09)0.53 (0.10)0.48 (0.08)0.49 (0.09)0.51 (0.09)
cQUtA (mL/min)550 (263)453 (233)439 (159)486 (233)429 (194)291 (154)572 (261)

We observed a statistically significant correlation (R = 0.76; P < 0.0001) between mQUtA and cQUtA (Figure 2). Table 2 shows the predicted values of mQUtA based on the given values of cQUtA, with various prediction intervals. The difference between QUtA measurements obtained by the two independent methods increased with increasing blood flow rate. When 31 pairs of log-transformed QUtA measurements were tested with Bland–Altman analysis, the geometric mean of the ratio cQUtA/mQUtA was found to be 1.14 (i.e. cQUtA exceeded mQUtA by 1.14 times on average) with 95% limits of agreement of 0.57–2.27 (Figure 3). The mQUtA correlated positively with Doppler-derived uterine artery PSV (R = 0.50; P = 0.004), EDV (R = 0.53; P = 0.002), TAMXV (R = 0.69; P < 0.0001) and TAMEANV (R = 0.75; P < 0.0001). The spatial velocity distribution coefficient for the uterine artery, calculated as a ratio between TAMEANV and TAMXV, was 0.57 (95% CI, 0.53–0.60).

thumbnail image

Figure 2. Correlation between Doppler-derived uterine artery volume blood flow (cQUtA) and uterine artery volume blood flow measured directly using a perivascular transit-time flow probe (mQUtA), showing the regression line and its 95% prediction intervals. Equation for the regression line: mQUtA = 79.08 + cQUtA × 0.74; R = 0.76; P < 0.0001. equation image, baseline; equation image, noradrenaline; equation image, beta-blocker; equation image, hypoxemia; equation image, recovery; equation image, hypotension; equation image, ephedrine.

Download figure to PowerPoint

thumbnail image

Figure 3. Bland–Altman plot demonstrating agreement between uterine artery volume blood flows (QUtA) measured non-invasively (cQUtA) and invasively (mQUtA), with 95% limits of agreement. Mean QUtA = (cQUtA + mQUtA)/2.

Download figure to PowerPoint

Table 2. Predicted values of uterine artery volume blood flow measured directly using a perivascular transit-time flow probe (mQUtA), based on values of Doppler-derived uterine artery volume blood flow (cQUtA)
cQUtAmQUtAPrediction interval
(mL/min)(mL/min)80%90%95%
100153.1− 33.2 to 339.4− 86.1 to 392.2− 131.9 to 438.1
200227.144.6 to 409.6− 7.1 to 461.3− 52.0 to 506.2
300301.1121.3 to 480.970.3 to 531.926.1 to 576.1
400375.1196.7 to 553.5146.2 to 604.1102.3 to 648.0
500449.1270.9 to 627.4220.3 to 678.0176.5 to 721.8
600523.2343.7 to 702.6292.8 to 753.5248.7 to 797.6
700597.2415.3 to 779.1363.7 to 830.6319.0 to 875.3
800671.2485.7 to 856.7433.1 to 909.3387.5 to 954.9
900745.2554.9 to 935.5501.0 to 989.5454.2 to 1036.3

Discussion

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

Volume blood flow analysis is considered to be the most accurate way of assessing organ or tissue perfusion. However, non-invasive measurement of volume blood flow is associated with considerable problems2, 3, especially when the flow is not laminar and the diameter of the blood vessel is small. Additionally, the spatial mean velocity in the arteries changes throughout the cardiac cycle, which makes its determination difficult. However, with most current sophisticated ultrasound equipment, measuring TAMEANV is possible13.

Doppler blood flow velocity measurements are repeatable if a vessel can be located and examined properly14. In pregnant women, the uterine arteries are identified relatively easily using color Doppler and the flow velocity waveforms can be obtained using pulsed-wave Doppler. In contrast to umbilical and fetal blood vessels, uterine arteries do not change their position during examination and can be insonated at a standardized location with an acceptable angle. However, it is difficult to measure accurately the diameter of uterine arteries using two-dimensional transabdominal ultrasonography. Although conventional color Doppler demonstrates reliably the blood vessel and the direction of flow, it does not delineate accurately the vessel diameter due to blooming (extension of color beyond the vessel wall). The use of power Doppler (amplitude mode) angiography to measure the diameter of uterine arteries has been suggested to be more reliable4, 5, but proper validation of the technique is still lacking. As the errors in diameter measurement can be substantial, whether other methods of delineating the blood vessels, such as B-flow imaging, would yield more accurate results merit further investigation.

The main purpose of our study was to compare the accuracy of measurement of QUtA by a non-invasive Doppler technique with measurement of volume blood flow directly using a transit-time ultrasonic perivascular flow probe. Our study showed that QUtA measured non-invasively correlates well with volume blood flow measured directly, validating its use for clinical purposes. However, the agreement between invasive and non-invasive methods of blood flow measurement was not perfect. Yet, even with this level of accuracy, the non-invasive measurement of QUtA may provide clinically relevant information, especially when volume blood flow assessment is performed longitudinally, although this needs to be further evaluated in clinical studies. Furthermore, there were no significant changes in the diameter of the uterine artery during the different experimental phases, whereas uterine artery absolute velocities (especially TAMEANV and TAMXV) correlated well with mQUtA, suggesting that they could be used as surrogate measures of QUtA.

Considerable expertise is required for the accurate measurement of TAMEANV. Additionally, not all commercially available ultrasound systems have the software capability to perform this measurement. Therefore, use of TAMXV rather than TAMEANV, with correction for velocity profile, may be considered more appropriate. However, accurate knowledge of spatial velocity profiles is necessary for this purpose. The spatial velocity distribution coefficient that characterizes the velocity profile in a vessel can be calculated as a ratio between TAMEANV and TAMXV15. In the pregnant sheep, we found this to be approximately 0.6 for the uterine artery, similar to the value reported for the umbilical artery in human fetuses16. However, variations in the vessel curvature and the distance from the insonation site to the bifurcation are likely to have an influence on the velocity profile of the uterine arteries. Thus, it may be preferable to use TAMEANV when possible, which is not dependent on a known velocity profile coefficient.

We have shown that QUtA can be measured non-invasively with reasonable accuracy, which makes it possible to estimate uterine vascular resistance using maternal BP and uterine volume blood flow measured non-invasively. Regarding the validity of our measurements, we used QUtA measured by a perivascular ultrasonic transit-time flow probe as the gold standard against which we compared the volume blood flow obtained non-invasively using Doppler ultrasonography. This method has been well validated in vitro and in vivo17, 18, with a reported variability of between 1.1 and 4.4% for volume blood flow measured by 6-mm probe17, and the manufacturer's website (www.transonic.com) states an absolute accuracy of ± 10% and a relative accuracy of ± 2%. Furthermore, unlike electromagnetic flow probes, the ultrasonic transit-time flow probe does not need to be in direct contact with the vessel wall (i.e. it does not need to fit the vessel size accurately), it is easy to calibrate and blood flow measurements are not affected by the hematocrit and vessel wall thickness17.

In summary, QUtA measured non-invasively by Doppler ultrasonography correlates well with volume blood flow measured directly. Uterine artery absolute blood flow velocities reflect uteroplacental volume blood flow in pregnant sheep.

Acknowledgements

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

We would like to thank Tom Wilsgaard, PhD, Associate professor, Institute of Community Medicine, University of Tromsø, Norway for his help with data analysis. Financial support for this study was provided by the Academy of Finland and University Hospital of Northern Norway. Prof. Huhta is supported by the Daicoff-Andrews Chair in Perinatal Cardiology of the University of South Florida and All Children's Hospital Foundations.

References

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