Chapter 10: DOPPLER VELOCIMETRY OF THE UTEROPLACENTAL CIRCULATION

The Doppler effect is named after Johann Christian Andreas Doppler (1803 to 1853), an Austrian scientist who described the changing color of stars as they move closer or farther away from the earth. Doppler signals, traveling back towards the ultrasound transducer which emitted the sound, are composed of many frequency shifts, thus creating a complex waveform.

Doppler velocimetry of the uteroplacental circulation is an important modality due to several reasons. It is a noninvasive diagnostic tool that provides insight and can assess resistance to blood flow in this circulation. It has been proposed as a screening test to identify women at high risk for the development of preeclampsia and small-for-gestational-age (SGA) fetuses. In addition, uterine artery Doppler has increasingly been used to screen for these disorders as early as the first trimester.

In this chapter, we review the anatomy and development of the uteroplacental circulation, the pathologic changes associated with placental implantation failure, the rationale for using Doppler velocimetry to detect these changes, and the studies that examine the effectiveness of uterine artery Doppler velocimetry in the first and second trimesters to screen for adverse perinatal outcomes, such as preeclampsia and SGA.

Blood supply to the uterus is mainly provided by the uterine arteries, with a small contribution from the ovarian arteries. The uterine arteries are branches of the internal iliac artery. Upon reaching the isthmic portion of the uterus, they ascend through the lateral wall before anastomosing with the ovarian arteries at the cornu of the uterus (Figure 10-1).¹ Blood supply to the anterior and posterior walls is provided by the arcuate arteries, which run circumferentially around the uterus. Radial branches extend from the arcuate arteries at right angles towards the endometrium, where they divide into 2 or more spiral arteries (Figure 10-2).^{2,3,4,5, and 6}

During pregnancy, approximately 100 spiral arteries connect the maternal circulation to the intervillous space of the placenta.² These vessels undergo important physiologic modifications to accommodate the 10-fold increase in blood flow necessary to meet the metabolic requirements of the fetus and placenta. The conversion of the spiral arteries to uteroplacental arteries is called "physiologic change." During the first trimester, a first wave of endovascular trophoblast invades the walls of the spiral arteries in the decidua, stopping at the decidual–myometrial junction at approximately 15 weeks. Subsequently, during the second trimester, a second wave of endovascular trophoblast invades the myometrial segment of the spiral arteries, interacts with the terminal segment of the radial arteries, and replaces their musculoelastic wall tissue with a mixture of fibrinoid and fibrous tissue.³ This second phase occurs at 14 to 24 weeks' gestation. Therefore, due to the conversion of small muscular spiral arteries into large vascular channels, there is transformation of the uteroplacental circulation from a high-resistance to a low-resistance vascular system.

In pregnancies complicated by preeclampsia and/or SGA fetuses, however, trophoblastic invasion is almost completely restricted to the decidual segment of the spiral arteries, with little or no evidence of invasion beyond the decidual–myometrial junction. Failure of normal trophoblastic invasion beyond 24 to 26 weeks' gestation results in high impedance to blood flow in the uteroplacental circulation, and is associated with the development of preeclampsia later in pregnancy. Therefore, Doppler velocimetry of the uterine arteries has the potential to detect high impedance to blood flow and, accordingly, to identify patients at risk for the development of preeclampsia and SGA fetuses.^{4,5,6,7,8,9, and 10}

Both continuous-wave (CW) and pulsed-wave (PW) Doppler have been employed in the study of the uteroplacental circulation. Although CW Doppler has the advantage of being less expensive, it does not allow visualization of the vessels being sampled. Therefore, the waveform must be identified by pattern recognition (also known as audio vessel signature), which raises concerns about the reproducibility of the method. This issue was addressed in a study of 35 normal pregnancies between 20 and 40 weeks, in which the right uterine artery was evaluated with CW and PW Doppler by the same investigator, to reduce interobserver variation.¹¹ No difference was observed between systolic-to-diastolic (S/D) ratios obtained with either CW or PW Doppler (2.03 ± 0.67 versus 2.03 ± 0.46, respectively; P >.10). Although the correlation between the 2 methods was low, it was statistically significant (r = 0.58, P <.01). However, when uterine artery S/D ratios were elevated (>3.0), the indices obtained with CW Doppler were significantly higher than those obtained with PW Doppler, suggesting that pattern recognition may be less reliable when the uterine artery S/D ratio is elevated. In this circumstance, the waveform may appear similar to, and be confused with that of, the iliac artery signal.

Color Doppler imaging allows accurate real-time mapping of the uterine arteries as they cross the external iliac arteries, proper placement of the pulsed Doppler gate, and is the preferred technique for performing uterine artery Doppler velocimetry.¹² When the examination is conducted during the second trimester, a general obstetrical ultrasound is performed first and the placental location should be determined. The transducer is then directed toward the iliac fossa, the external iliac artery is imaged in a longitudinal section, and the uterine artery mapped with color Doppler as it crosses the external iliac artery (Figure 10-3). This procedure allows precise positioning of the pulsed Doppler gate within the vessel lumen and acquisition of appropriate waveforms for spectral analysis (Figure 10-4). Transvaginal ultrasound is used for studies performed during the first trimester (Figure 10-5). Close proximity to the pelvic structures, the use of high-frequency transducers, and a low angle of insonation all improve image quality and Doppler signals.

Measurements obtained from the Doppler velocity waveform allow the calculation of angle-independent Doppler indices that reflect the degree of resistance to blood flow downstream of the interrogation site: (1) systolic/ diastolic ratio (S/D), (2) resistance index (RI), and (3) pulsatility index (PI). Defective trophoblast invasion and failure of physiologic transformation of uterine vessels results in persistent uterine artery notching and/or elevation of the uterine artery Doppler indices. Vascular resistance is estimated by comparing systolic and diastolic waveforms. Normally, the uterine vascular bed is a low-resistance circuit, and flow continues throughout diastole. However, as resistance increases, diastolic velocity diminishes in relation to systolic flow. Uterine artery abnormality has been defined as either an absolute cutoff (for example, RI >0.58) or a measurement greater than a particular centile of the reference range. Studies using receiver operator characteristic or ROC curves have shown a PI greater than 1.5,¹³ an RI greater than 0.68,¹⁴ or an RI in the region of the 90th centile¹⁵ to predict the development of preeclampsia and SGA.

The pulsatility index is currently the most commonly used index for the evaluation of uterine artery Doppler waveform patterns.¹⁶ In general, there are several advantages to utilizing the PI for analyzing both arterial and venous Doppler flow velocity waveforms. The PI describes the shape of the velocity waveform much better, since it includes the area below the curve in the formula, irrespective of qualitative changes. For example, the presence of uterine artery notching will be reflected in the time averaged velocity, and accordingly the PI; however, this will be missed by the S/D ratio or RI. In addition, there are inherent problems with the S/D ratio, as these tend to approach infinity as the D approaches zero. Thus, at low end-diastolic velocities, when the PI is used (instead of S/D ratio), there will be a reduction in the margin of error for the measurement.

Once the uterine artery waveform is obtained, it can also be subjectively inspected for the presence of an early diastolic notch (Figure 10-6), which is associated with the subsequent development of preeclampsia.^{17,18,19,20, and 21} An early diastolic notch is defined as a persistent decrease in blood flow velocity in early diastole, below the diastolic peak velocity.¹⁶ The diastolic notch is characteristic of vessels having resistance, and is thought to represent the elasticity of the vessel.²² Although good interobserver agreement has been demonstrated for the subjective assessment of diastolic notching among experienced operators,^19,23 newer Doppler indices have been proposed to objectively assess the shape of the uterine artery waveform, in particular the presence and intensity of the early diastolic notch (Figure 10-7).^13,24,25 These indices are the peak systolic over protodiastolic ratio (A/C)^13,24,25 and the maximum diastolic velocity minus protodiastolic velocity over end-diastolic velocity ratio (D – C/B) (Figures 10-8 and 10-9).¹³ Irion et al²⁵ reported that an A/C ratio equal to or greater than 2.5 had a sensitivity of 88%, and a specificity of 86% to detect a notch. In a subsequent study comparing objective and subjective assessments of 50 abnormal uterine artery Doppler waveforms, Bower et al¹³ found that a (D – C/B) ratio greater than 0.15 reached the closest agreement with the subjective diagnosis of a diastolic notch.

Others have proposed a different method to quantify the diastolic notch of the uterine artery waveform (Notch Index), and it is intended to remove the subjectivity of defining a notch.²⁶ The Notch Index is derived from the difference between the peak velocity in diastole and the velocity at the nadir of the notch, divided by the mean velocity over 1 cycle length. However, when compared to the PI, the Notch Index at 21 to 22 weeks' gestation did not lead to better predictive values for hypertensive disorders of pregnancy.²⁶

Recently, a prospective, cross-sectional study of 620 women was conducted to create reference ranges for uterine artery mean PI from 11 to 41 weeks' gestation (20 patients for each week).¹⁶ The PI of the left and right arteries were measured, and the mean PI was then calculated. The presence or absence of a bilateral early protodiastolic notch was also recorded. The prevalence of bilateral notching was 46% (11 to 14 weeks), 16.5% (15 to 24 weeks), and 5% (25 to 41 weeks), which shows that the prevalence of bilateral notching decreases with advancing gestational age. There was also a significant decrease in the mean PI between 11 weeks (1.79; 95th centile, 2.70) and 34 weeks (0.70; 95th centile, 0.99). The mean PI then became more stable up until 41 weeks (0.65; 95th centile, 0.89).

A recent study by Lau et al has reported for the first time reversed diastolic flow in the uterine artery.²⁷ This occurred in 2 obstetric cases that were complicated by severe placental insufficiency. In the first case, uterine artery Doppler interrogation at 29 weeks showed high PI and RI, along with notching of the left uterine artery, while reversed diastolic flow was noted in the right uterine artery. At 30 weeks' gestation, the patient developed severe preeclampsia and then developed eclampsia postpartum. In the second case, fetal growth restriction became apparent from 26 weeks onwards. At 29.6 weeks' gestation, reversed diastolic flow was noticed in the right uterine artery, and the patient developed proteinuria with edema. Abnormal cardiotocographic findings were seen, and an emergency cesarean section was performed. The authors hypothesized that reversed diastolic flow in the uterine artery in the third trimester is the end result of progressive deterioration in the underlying vasculopathy of placental insufficiency.

Nonpregnant Uterus

Uterine artery waveforms obtained from the nonpregnant uterus are characterized by high impedance to blood flow. Doppler indices (S/D ratio, PI, and RI) also change according to the phase of the menstrual cycle.²⁸ Absent end-diastolic velocities and early diastolic notches are more pronounced during the follicular phase.^29,30 In a study of 150 normal women, Kurjak et al³¹ reported the average uterine artery RI during the proliferative phase to be 0.88 ± 0.04 (2 SE). The RI started to drop 1 day before ovulation, reached a nadir of 0.84 ± 0.04 (2 SE) on day 18, and remained at this level for the rest of the cycle.

High resistance to flow in the uterine artery during the midluteal phase of the cycle (day 21) has been associated with infertility.^32,33 Steer et al³³ studied a group of 23 women having artificial insemination with donor sperm because their partners were azoospermic, and also 161 women receiving treatment for infertility (35 unexplained infertility, 91 with tubal disease, 8 with endometriosis, and 22 with anovulatory infertility). Table 10-1 shows a comparison between the uterine artery pulsatility indices obtained during the midluteal phase of the menstrual cycle for each group. Regardless of specific etiology, the PI was higher in women with infertility.

With regards to recurrent miscarriage, abnormal uterine perfusion is also considered an etiological factor of this condition. In women undergoing in vitro fertilization, those with a higher PI on the day of follicular aspiration had a lower probability of successful pregnancy.^34,35 Steer et al³⁶ demonstrated that a mean PI greater than 3.0 before embryonic transfer predicted up to 35% of implantation/pregnancy failures. These findings suggest a potential value for uterine artery Doppler velocimetry in identifying endometrial receptivity before in vitro fertilization. It seems that in cycles having a high uterine impedance, embryos could be cryopreserved until the uterus becomes more receptive.

Similarly to patients with infertility, recent studies have also demonstrated in patients with a history of recurrent miscarriage an increased resistance to uterine artery blood flow in the midluteal phase of the cycle that is higher with respect to that found in normal, fertile patients.^37,38 Ferreira et al found that women with a history of recurrent pregnancy loss had a significantly higher uterine artery PI (in second phase of menstrual cycle) than those in the control group (2.71 ± 0.54 versus 2.30 ± 0.44, respectively).³⁸ The mechanisms by which abnormal uterine perfusion determines pregnancy loss are not completely understood. It has been postulated that impaired uterine perfusion could be related to an underlying subclinical vasculopathy. This has been demonstrated by increased levels of substances (adrenomedullin, which mediates vasodilatory properties) produced in response to vascular damage, found in patients with recurrent miscarriage and abnormal uterine perfusion.³⁹ Therefore, it has been suggested that vasoactive substances (enhancing uterine perfusion), may effectively improve the possibility for a successful pregnancy. In accordance with this hypothesis, several therapeutic approaches have been proposed to improve uterine perfusion in these patients, including treatment with nitric oxide donor (NO), sildenafil, and low-dose aspirin.^{40,41,42, and 43} A recent study also found that low-dose aspirin and omega-3 fatty acids were effective in improving uterine artery blood flow velocity in women with recurrent miscarriage due to abnormal uterine perfusion.⁴⁴

First Trimester

In human pregnancies, the intervillous maternal circulation is only fully established at approximately 11 to 12 weeks.^45,46 Before 12 weeks, the intervillous space of the definitive placenta is separated from the uterine circulation by a trophoblastic shell, and trophoblastic plugs obliterate the tip of the uteroplacental arteries. Thus, a very limited amount of maternal blood reaches the intervillous space during the first trimester, and color Doppler signals are usually not seen in the placental substance (Figure 10-10).^{47,48,49,50, and 51} This is consistent with the clinical observation that chorionic villus samples obtained before 12 weeks are rarely tinted by blood, in comparison with samples obtained in the second trimester or later.⁴⁶ Detection of blood flow in the intervillous space during the first trimester is associated with higher pregnancy loss rates (Figure 10-11).^52,53 This increased circulation may conceivably lead to excess pressure, which is probably followed by villous capillary collapse and arrest of the embryonic placental circulation, provoking the premature separation of the gestational sac.⁵⁴ After 12 weeks' gestation, blood flow can be systematically detected by color Doppler imaging of the placental substance.^{47,48,49,50, and 51}

In contrast, Doppler flow velocity waveforms can be obtained from the implantation site as early as 5 weeks after the last menstrual period, and are characterized by low impedance to blood flow. A prominent diastolic component is observed, with the mean RI ranging from 0.41 ± 0.10 (SD) to 0.48 ± 0.08 (SD).^47,55,56 The RI gradually decreases from 6 to 12 weeks.⁵⁷ Flow velocity waveforms obtained from the uterine arteries are characterized by elevated S/D ratios (Figure 10-12), reduced end-diastolic velocities, and the presence of a notch in the systolic deceleration slope.⁵⁷ The vascular resistance decreases as gestation progresses into the second trimester (Figure 10-13).^{57,58,59,60, and 61}

Coppens et al⁵¹ conducted a longitudinal study of uteroplacental blood flow in 37 normal early human pregnancies. Uterine artery blood flow was characterized by a high systolic component and low end-diastolic flow pattern, associated with a prominent early diastolic notch, consistently present from 8 to 14 weeks of gestation (see Figure 10-5). Waveforms from the arcuate arteries also showed a diastolic notch but had higher diastolic velocities in comparison to the uterine arteries. The spiral arteries showed the lowest resistance, with high diastolic blood flow and a small diastolic notch that disappeared in about half of the cases by 10 weeks, and in all cases by 13 weeks. In 62% of the cases, the diastolic notch of the arcuate arteries disappeared the week after the disappearance of the notch in the spiral arteries. A recent article that reviewed Doppler studies of uterine arteries in normal first trimester pregnancies found that most of the reported studies described an early diastolic notch in the uterine artery Doppler waveforms (suggesting a high vascular resistance) that will progressively disappear during the second trimester as a consequence of the decreased resistance downstream.⁶²

Normal mean values (±2 SD) for PI in uterine, arcuate, and spiral arteries (from 8 to 14 weeks) are presented in Table 10-2.

Second and Third Trimesters

The second trimester is characterized by a progressive enlargement of the cross-sectional area of the main uterine artery,^59,63 increase in peak velocity and volume flow rates,^{59,60, and 61,64,65, and 66} and a progressive fall in impedance to blood flow.^{57,58,61,64,67,68, and 69} The physiological process of trophoblastic invasion is reflected in the observation from Doppler studies that impedance to flow in the uterine arteries decreases with gestation between 6 and 24 weeks, and remains constant thereafter.⁷⁰ The initial fall until 24 to 26 weeks is believed to be due to trophoblastic invasion of the spiral arteries, but a continuing fall in impedance may also partly be explained by a persistent hormonal effect on arterial wall elasticity. The diastolic notch and the difference between uterine artery S/D ratios of placental versus nonplacental sites should disappear after 24 to 26 weeks of gestation.^{58,59,61,64,71,72} Normal values for the RI of the uterine and arcuate arteries (placental versus nonplacental sites) from 16 to 26 weeks' gestation are shown in Figures 10-14, and 10-15.

Puerperium

Within a few hours after delivery, profound modifications in uterine perfusion occur. In the first 24 hours postpartum, the blood flow falls considerably, the S/D ratio increases, and the blood velocity remains high. By the second day postpartum, both the S/D ratio and PI increase considerably, and the diastolic notch reappears. After an initial rise in impedance to blood flow, no further changes are noted until the 6th puerperal week, when the RI starts to rise again, lasting until the third month of puerperium.^67,72

Influence of Sampling Site and Placental Implantation on Doppler Flow Velocity Waveforms

Choosing the appropriate vessel to obtain the Doppler sample is important. Figure 10-16 shows sampling sites used by different investigators to obtain uteroplacental Doppler flow velocity waveforms. These sites include the main uterine arteries,⁷³ the arcuate arteries along the lateral uterine wall,⁷⁴ and the subplacental vessels.^69,75 Much controversy has been generated by comparing studies using different sampling sites.⁷⁶ Studies evaluating the arcuate, radial, and spiral arteries during the second trimester are subject to sampling bias and are difficult to reproduce. Furthermore, Doppler waveforms obtained from these sites do not necessarily reflect the status of the whole uteroplacental circulation, because only a small section is examined. On the other hand, Doppler waveforms from the main uterine arteries provide a more accurate picture of perfusion status, reflecting the sum of resistances in the uteroplacental vascular bed.^{77,78, and 79} As mentioned before, color Doppler imaging allows easy identification of the main uterine artery as it crosses the external iliac artery, thus improving reproducibility (see Figure 10-3). Waveforms obtained from this site are comparable to waveforms obtained from the main uterine artery near its origin from the internal iliac artery.⁷⁸

The combination of RIs from 4 different locations (right and left uterine and arcuate arteries) into an averaged index has been proposed as a means to overcome methodologic problems when comparing studies of Doppler waveforms that were sampled at different sites.⁸⁰ However, the predictive values for abnormal pregnancy and perinatal outcomes were lower than those reported by studies using the uterine or arcuate arteries alone. Therefore, this proposed approach is not justified and does not warrant substitution by this method.

Sampling both uterine arteries is also important.⁸¹ There are differences in the Doppler indices of the placental and nonplacental sides. The S/D ratios, RI, and PI are lower in the ipsilateral site of placental implantation.^32,58,82 This difference is more pronounced during the first half of the pregnancy, and should gradually disappear as the pregnancy approaches the third trimester (see Figures 10-14 and 10-15).^32,58,78 The lower impedance in the uterine artery on the same site as the placenta is thought to be attributed to the trophoblastic invasion only taking place in placental spiral arteries, and the resulting fall in impedance being transmitted via collaterals to other parts of the uterine circulation. In fact, it has been shown that an abnormal uterine artery waveform on the placental side is a better predictor of hypertensive disease than one on the nonplacental side.⁸³ If the placenta is not preferably implanted on one side, both uterine arteries should have similar indices.

In a study of 71 pregnancies, Schulman et al⁷³ demonstrated that persistence of a difference between the S/D ratios of left and right uterine arteries after 26 weeks of gestation (Δ ≥1) was associated with an increased risk for the development of preeclampsia (50% versus 14%; relative risk [RR] = 2.55, 95% confidence interval [CI] = 1.5 to 4.3), and SGA [39% versus 5%; RR = 2.88, 95% CI = 1.8 to 4.6].

Elevated Doppler Indices in the Uterine Arteries as a Measure of Impedance to Blood Flow in the Uteroplacental Circulation

Evidence that elevated S/D ratios, RI, and PI, and the presence of a diastolic notch in the uterine artery flow velocity waveform reflects increased impedance to blood flow in the placental circulation comes from 3 sources: computer models of the uteroplacental circulation,^84,85 embolization of the uteroplacental circulation in animal models,^86,87 and studies correlating histologic evaluation of placental bed biopsies with Doppler velocimetry of the uterine arteries.^88,89

Computer Models of the Uteroplacental Circulation

The uterine artery flow velocity waveform has 2 components: a pulsatile component and a steady flow component. The pulsatile component is formed by the interaction of an outgoing wave and a reflected wave, which bounces back to the maternal heart upon reaching the uteroplacental vascular bed.

Adamson et al⁸⁴ developed a computer model to evaluate the effect of placental resistance, arterial diameter, and blood pressure on the shape of the uterine artery velocity waveform. Their model was designed to simulate resistance to flow, vessel compliance, and main uterine artery blood inertia, as well as the physical properties of all the vessels branching from the main uterine artery to form the distal vascular bed (arcuate, radial, and spiral arteries). The variables analyzed (uteroplacental vascular resistance, uterine artery radius, and mean arterial blood pressure or MAP) were altered over realistic ranges, either independently or in combination, to examine their effects on the uterine artery flow velocity waveform. When the resistance of the distal vascular bed was increased 6-fold, while keeping the uterine artery radius and MAP constant (Figure 10-17), the following observations were made: (1) decreased mean velocity and blood flow; (2) elevated PI and S/D ratios; and (3) development of a notch in the early diastolic portion of the flow velocity waveform, caused by destructive interaction between outgoing and reflected waves.

A reduction in the uterine artery radius resulted in an increase in mean velocity, elevated PI and S/D ratios, but no diastolic notch. An increase in the MAP caused elevation in mean blood velocity without changing the RI, or generating a diastolic notch. Importantly, these observations suggest that changes in uterine artery Doppler velocimetry observed in preeclampsia are not mediated by high blood pressure, but rather are the consequence of an abnormally elevated uteroplacental vascular resistance, expressed by elevations in Doppler indices and the presence of a diastolic notch.

Talbert⁸⁵ proposed another model to study uteroplacental blood flow. The model used resistance and capacitance elements in all components of the uteroplacental circulation, making it possible to simulate various physiologic situations and observe changes in the Doppler flow velocity waveforms. Uteroplacental blood flow parameters were changed within physiologic limits for a 1000-g human fetus at 28 weeks' gestation. The first experiment described was a 33% reduction in blood flow, achieved by raising the resistance to flow in the transcotyledonary region and spiral arteries. The results were increased uterine artery PI and RI, but no diastolic notch. In the second experiment, intervillous obstruction was modeled by increasing central cotyledon pressure by nearly doubling transcotyledonal resistance, which caused distention and reduction of flow resistance of the spiral and radial arteries. The uterine artery PI and RI increased, but no diastolic notch was produced. When placental resistances were kept within normal limits, and the distensibilities of uterine and arcuate artery walls were raised, a diastolic notch was produced, with normal uterine artery PI and RI. Thus, the typical flow velocity waveforms observed in vivo in abnormal uterine artery Doppler studies (increased PI, RI, and the presence of a diastolic notch) was produced by a combination of increased placental resistance and distensibility of the uterine and arcuate artery walls.

Embolization of the Uteroplacental Circulation of Pregnant Ewes

Ochi et al^86,87 embolized the uterine spiral arteries of pregnant ewes at 16 to 17 weeks of gestation through the stepwise injection of gelfoam microbeads into the uterine arteries, and then analyzed the associated Doppler velocimetric changes. Embolization of the uteroplacental circulation resulted in a dose-dependent decrease in blood flow, an increase in uterine vascular resistance, a dose-dependent attenuation of the pregnancy-related physiologic elevation in the diastolic flow velocity, an increase in PI,⁸⁶ and the appearance of a diastolic notch.⁸⁷ These findings suggest that elevated PI and the presence of a diastolic notch in the uterine artery velocity waveform indicate increased vascular resistance of the uteroplacental circulation.

Correlation of Placental Bed Biopsies and Uterine Artery Doppler Velocity Waveforms

Olofsson et al⁸⁸ and Lin et al⁸⁹ compared histologic findings of placental bed biopsies obtained at the time of cesarean section with uterine artery waveforms. They studied placental bed biopsies taken at cesarean section in 26 complicated (study group) and 29 uncomplicated (control group) pregnancies after performing Doppler velocimetry of the uterine arteries. Physiologic changes were absent in 77% (20 of 26) of the complicated pregnancies and present in all controls. Patients with abnormal placental bed biopsies were more likely to have elevated uterine artery PIs and abnormal pregnancy outcomes (pregnancy-induced hypertension or PIH, and SGA). Similar conclusions were reached by Lin et al,⁸⁹ who studied 43 patients having Doppler examination of the uterine arteries within 7 days of delivery and found that 92% (12 of 13) of the patients with a uterine artery S/D ratio greater than or equal to 2.5 failed to show trophoblast migration into the myometrium. Patients with impaired trophoblast migration into the myometrium also delivered more prematurely (33.1 ± 2.7 versus 36.1 ± 2.2 weeks, P <.02) and had a higher rate of SGA fetuses (1290.2 ± 319.1 versus 1870.5 ± 523.1 g, P <.02).

Prediction of Preeclampsia and SGA Fetuses

Abnormal uterine artery Doppler flow velocity waveforms are characterized by elevated S/D ratios, RI, or PI, and the presence of a diastolic notch (see Figure 10-6). As discussed before, pregnancies complicated by preeclampsia, and to a lesser extent SGA fetuses, show evidence of impaired trophoblastic invasion of the myometrial portion of the spiral arteries.

Therefore, many investigators have studied the potential role of uteroplacental Doppler to identify patients at risk for preeclampsia, SGA, preterm delivery, or adverse perinatal outcomes. Tables 10-3,10-4,10-5, and 10-6 summarize the results of such studies for high-risk (1987 to 1999; second trimester)^{14,90,91,92, and 93} or unselected populations (1986 to 1998; various gestational ages).^{94,95,96,97,98,99,100,101,102,103,104,105,106,107, and 108} In 14 studies,^{14,90,92,93, and 94,98,100,101, and 102,104,105,106,107, and 108} Doppler was considered effective in identifying a high-risk population for the development of preeclampsia and/or SGA; in 3,^91,95,96 the results were inconclusive; and in 3,^99,102,108 the results did not support a role for uteroplacental Doppler velocimetry as a screening test in pregnancy. Divergence among the studies is attributed to differences in patient selection and gestational age for screening, type of equipment used (CW, PW, or color Doppler), multiple definitions of abnormal flow velocity waveforms, different vessels examined, and heterogeneous outcome criteria.^14,109 In addition, there was a wide variation in the number of patients studied, and also in the prevalence of adverse outcomes.

Sensitivities for the detection of preeclampsia ranged from 25% to 100%. The sensitivity improved as gestational age approached 26 weeks, when the main uterine arteries were the vessels interrogated (rather than other vessels, such as arcuate arteries), and when a persistent diastolic notch was one of the criteria used for analysis.¹⁰² Positive predictive values (PPVs) for preeclampsia ranged from 2% to 50% in low-risk populations (see Table 10-5), and increased from 17% to 70% in high-risk populations (see Table 10-3). Sensitivities for the detection of SGA in low-risk populations ranged from 0% to 87%, whereas the PPV ranged from 0% to 54%.

Albaiges et al performed one-stage screening for pregnancy complications (preeclampsia, fetal growth restriction, fetal death, and placental abruption) by utilizing Doppler assessment of the uterine arteries at 23 weeks' gestation in 1757 pregnancies.¹¹⁰ The sensitivities of increased mean uterine artery PI or bilateral notches were as follows: (1) 45% (95% CI = 32.2 to 57.4) for preeclampsia; (2) 90% (95% CI = 55.5 to 99.7) for preeclampsia that required delivery before 34 weeks; (3) 70% (95% CI = 34.8 to 93.3) of those whose infants had birth weights less than the 10th percentile and were delivered before 34 weeks; (4) 50% (95% CI = 18.8 to 81.3) of those with placental abruption that required emergency delivery; and (5) 83% (95% CI = 35.9 to 99.6) of fetal deaths. With regard to PPV, women at highest risk were those with both a high mean uterine artery PI and bilateral notches. These women had a 40% chance of developing preeclampsia, and a 45% chance for delivering infants weighing less than 10th percentile. Another important finding was the negative predictive value (NPV), which was more than 99% for adverse outcomes before 34 weeks, placental abruption, and fetal death.¹¹⁰ This suggests that women with normal uterine artery Doppler results are unlikely to develop preeclampsia, fetal growth restriction, or placental abruption, and therefore do not necessarily need antenatal follow-up that is as close as that required in women with abnormal uterine artery Doppler findings. In conclusion, a one-stage uterine artery screening program at 23 weeks identified most women who subsequently developed serious complications of impaired placentation associated with delivery before 34 weeks.

A large, multicenter study in 7851 patients was conducted by the Fetal Medicine Foundation to screen for preeclampsia and fetal growth restriction by performing uterine artery Dopplers at 23 weeks.¹¹¹ The mean PI of the 2 arteries was calculated and the presence of an early diastolic notch was noted. The sensitivity of mean uterine artery PI greater than 95th centile (false-positive rate or FPR, 5%) was (1) 69% for preeclampsia (with fetal growth restriction); (2) 24% for preeclampsia (without fetal growth restriction); and (3) 13% for fetal growth restriction (without preeclampsia). However, when delivery occurred before 32 weeks, the respective sensitivities increased to 93%, 80%, and 56%, respectively. The sensitivity of bilateral notches in predicting preeclampsia and/or fetal growth restriction was similar to that of increased PI, but the screen-positive rate with notches (9%) was much higher than that with increased PI (5%). Therefore, an important point is that the sensitivity for both preeclampsia and fetal growth restriction increases with the severity of the disease (as defined by earlier gestational age at delivery), or is inversely related to the gestational age at delivery. This study showed that a one-stage Doppler screening program at 23 weeks' gestation will identify most women who subsequently develop severe (and thus more clinically relevant) preeclampsia and/or fetal growth restriction.

A recent study by Yu et al performed a prospective screening study for preeclampsia using uterine artery Doppler ultrasound in unselected, low-risk singleton pregnancies (32,000 women).¹¹² There were 612 cases of preeclampsia, and 144 required early delivery (<34 weeks). When performing uterine artery Dopplers at 22 weeks' gestation either alone or in combination with maternal history, detection rates were 52% and 57%, respectively (FPR 10%). More importantly, Doppler was found to be particularly effective in screening for severe preeclampsia that necessitated iatrogenic delivery before 34 weeks, with a detection rate of 85% (FPR 10%).

In 2008, a large, prospective, multicenter study (30,639 pregnancies) from the Fetal Medicine Foundation investigated the relationship between preeclampsia, SGA and gestational age at delivery, and the effect of this relationship on the prediction of preeclampsia via uterine artery Doppler imaging.¹¹³ Doppler of the uterine artery was performed at 22 to 24 weeks' gestation in unselected women with singleton pregnancies. In 2% of cases, women developed preeclampsia, and in this group there was an inverse significant association between gestational age at delivery and prevalence of SGA, and between gestational age at delivery and mean uterine artery PI and prevalence of mean uterine artery PI greater than 95th percentile. The mean uterine artery PI was greater than 95th percentile in (1) 77% of women who developed preeclampsia (requiring delivery before 34 weeks); (2) 36% of those delivering at 34 to 37 weeks; and (3) 22% of those delivering after 37 weeks. The respective percentages were higher for those with preeclampsia and SGA, but lower for those with SGA (but without preeclampsia). Similarly to others, the authors concluded that uterine artery Doppler assessment is more effective in identifying preeclampsia requiring preterm rather than term delivery.¹¹³

Bower et al¹⁹ proposed a two-stage screening protocol for preeclampsia with uterine artery Doppler velocimetry performed at 18 to 22, and 24 weeks' gestation. In their study of 2058 low-risk patients, 16% (n = 329) had abnormal flow velocity waveforms (defined as the worst RI above the 95th percentile of reference range, or an early diastolic notch in either uterine artery) at 18 to 22 weeks. The presence of a persistent notch in the early diastolic component of the waveform at 24 weeks improved the predictive value of the test for preeclampsia (from 12% to 28%) and was associated with a 68-fold increased relative risk for the development of preeclampsia. Importantly, all women who delivered before 34 weeks because of severe preeclampsia had abnormal waveforms at both stages of screening. The authors concluded that preeclampsia can be predicted effectively by two-stage Doppler screening.

Harrington et al¹⁰⁸ examined the possibility of predicting preeclampsia and SGA at 12 to 16 weeks with transvaginal uterine and umbilical artery Doppler velocimetry. The parameters evaluated included the presence or absence of an early diastolic notch, vessel diameter, RI, PI, time-averaged mean velocity (TAV), maximum systolic velocity and volume flow in the right and left uterine arteries, and RI and PI in the umbilical arteries. The main outcome measures of the study were intrauterine death, birth weight, preeclampsia, and antepartum hemorrhage. Women who subsequently developed preeclampsia had significantly lower mean uterine artery TAV, lower volume flow, and elevated RI. The odds ratio (OR) for the development of preeclampsia in women with bilateral notches was 43.54 (95% CI = 5.84 to 324.73). These patients were also more likely to have SGA babies (OR = 8.61, 95% CI = 4.0 to 20.0), or deliver prematurely (OR = 2.38, 95% CI = 1.19 to 4.75).

On the other hand, a large study conducted by Irion et al¹⁰⁷ in a low-risk population, with similar inclusion criteria, Doppler examination technique, and outcome variables as the study of Bower et al,¹⁹ was not able to reproduce the results described earlier. The likelihood ratio for the development of preeclampsia in patients with a notch in the uterine arteries examined at 26 weeks was only 1.94 (95% CI = 1.14 to 3.31). The authors concluded that Doppler of the uteroplacental circulation did not fulfill the requirements for a screening test in unselected populations. However, they acknowledged that the likelihood ratios for the abnormal test confirmed the interest of this technique in selected patients at higher risk of obstetrical or perinatal complications (ie, those with risk factors for preeclampsia identified by medical history, second trimester arterial blood pressure, or a combination of biological tests, as well as those with risk factors for intrauterine growth retardation, and prematurity).

Among the other studies that failed to show a benefit of uterine artery Doppler velocimetry as a screening test in pregnancy, the study by Newnham et al¹⁰² evaluated only the presence of SGA as an outcome variable. In the study of Hanretty et al,⁷⁵ PIs obtained from subplacental vessels (radial or spiral arteries) were compared between a group of 32 normal pregnancies and a group of 32 pregnancies complicated by preeclampsia. All examinations were performed during the third trimester, and waveforms with the lowest resistance to flow were selected for analysis. No differences in PI were detected between the 2 groups. However, because of a sampling bias, these results are not surprising. Placental bed biopsies from patients with preeclampsia show areas lacking trophoblastic invasion scattered among areas with normal physiologic change.⁶ By selecting only those areas with the lowest resistance to flow, the possibility of sampling vessels with normal trophoblastic invasion but with abnormal placentation is increased.

In 2004, Conde-Agudelo et al performed a World Health Organization (WHO) systematic review of screening tests for preeclampsia.¹¹⁴ Table 10-7 shows the results of updated meta-analyses of this review on the predictive accuracy of uterine artery Doppler velocimetry for preeclampsia.¹¹⁵ The majority were performed in the second trimester. Overall, it appears that the level of prediction of preeclampsia in low-risk populations is minimal to moderate, and in high-risk populations is minimal, regardless of the uterine artery Doppler indices used. In low-risk populations, the sensitivities and specificities of the uterine artery Doppler indices varied between 34% and 76%, and between 83% and 93%, respectively. The best predictors of preeclampsia were the presence of bilateral notches or an increased PI (positive likelihood ratio of 5). Although these indices appeared to have high specificity, this is at the expense of compromised sensitivity. For high-risk populations, all uterine artery Doppler indices had poor predictive ability, with positive likelihood ratios between 2.5 and 3.3, and negative likelihood ratios between 0.4 and 0.8. Combinations of notching and PI or RI did not improve the predictive accuracy of the single index.

By contrast, Table 10-8 presents the predictive accuracies of uterine artery Doppler velocimetry for early onset preeclampsia (≤34 weeks' gestation).¹¹⁵ With the exception of 1 study,¹⁷⁵ all were performed in the second trimester. Half of the studies reported positive likelihood ratios greater than 10, whereas the other half reported positive likelihood ratios between 5.0 and 10.0. Five studies reported negative likelihood ratios around 0.2. Therefore, regardless of the index or combinations of indices used, uterine artery Doppler velocimetry appears to be a moderate to good predictor for the development of preeclampsia appearing before 34 weeks' gestation.

When evaluating uterine artery Doppler studies, an important point that has been frequently overlooked is the high NPV of the test.^106,109 Studies in unselected populations have reported negative values ranging from 96% to 99%,^{94,95,96, and 97,101,102,104,105,108} and a substantial reduction in the risk of developing preeclampsia or delivering an SGA baby. Therefore, some have proposed that a normal uterine artery Doppler study at 24 weeks could serve to organize prenatal care by identifying a group of low-risk women who would be ideal candidates for community care during pregnancy.^106,109 Kurdi et al¹⁰⁶ performed uterine artery Doppler velocimetry in 946 low-risk women at 20 weeks to test this hypothesis, and found that 99% of the women with normal uterine artery Doppler flow velocity waveforms did not develop preeclampsia, and 96% did not deliver an SGA baby (birth weight below the 5th centile). Moreover, none of the women with normal Doppler studies required delivery before 37 weeks for preeclampsia or an SGA fetus. Indeed, in women with normal uterine artery Dopplers, the OR for developing any complication during pregnancy (defined as preeclampsia, placental abruption, delivery of an SGA baby below the 10th centile, or a pregnancy that resulted in stillbirth or a neonatal death) was 0.24 (95% CI = 0.17 to 0.34). The authors therefore concluded that these women form a "truly low-risk group, who should be suitable for community-led care, thereby enabling obstetricians to concentrate more time and energy on high-risk pregnancies."

In conclusion, uterine artery Doppler velocimetry meets some of the criteria for the ideal predictive test for preeclampsia (simple, rapid, inexpensive, noninvasive, innocuous).¹¹⁵ However, there has been a wide range of sensitivity and PPVs reported in predicting preeclampsia and SGA fetuses. Some studies have shown that uterine artery Doppler is considered effective in identifying a high-risk population for the development of preeclampsia and/or SGA. However, recent and updated meta-analyses have found that the level of prediction of preeclampsia in low-risk populations is minimal to moderate, whereas for high-risk populations it is minimal, regardless of the Doppler indices used.¹¹⁵ Accordingly, the conclusion from the meta-analyses is that current evidence does not support the routine use of uterine artery Doppler for the prediction of preeclampsia in clinical settings.¹¹⁵ Importantly, however, because studies have shown that second trimester uterine artery Doppler is better (moderate to good) at predicting the development of preeclampsia before 34 weeks (early onset) (see Table 10-8), this method could be beneficial in the prediction of this condition. Finally, uterine artery Doppler screening has a high NPV (substantial reduction in the risk of developing preeclampsia or delivering an SGA baby).

Prediction of Superimposed Preeclampsia in Patients With Chronic Hypertension

The incidence of superimposed preeclampsia in pregnancies complicated by chronic hypertension ranges from 4.7% to 52%, depending on the severity of hypertension at the onset of pregnancy, and on the diagnostic criteria used.¹⁷⁸ In this group of patients, superimposed preeclampsia is strongly associated with a worse perinatal outcome.¹⁷⁸ It is no wonder, then, that uterine artery Doppler velocimetry has been investigated as a test to predict those who could develop superimposed preeclampsia among patients with chronic hypertension.^179,180

Caruso et al¹⁷⁹ studied 42 women with chronic hypertension who had uterine artery Doppler velocimetry performed at 23 to 24 weeks. The prevalence of superimposed preeclampsia was 21%, and the lowest uterine artery waveform RI greater than 90th centile predicted this outcome with a sensitivity of 100%, specificity of 88%, PPV of 69%, and NPV of 100%. Patients with abnormal uterine artery Doppler velocimetry were also more likely to deliver a premature (30.5 ± 3.3 versus 38.3 ± 1.8 weeks, P = .0005) or low-birth-weight baby (1140 ± 651 versus 3034 ± 532 g, P = .0005), have a cesarean section for fetal distress (46% versus 0%, P = .0005), deliver a newborn with an Apgar score below 7 at 5 minutes (61.5% versus 0%, P = .0005), or have their baby admitted to the neonatal intensive care unit (69.2% versus 13.8%, P = .0005). Perinatal mortality rate was also higher for patients with abnormal uterine artery Doppler velocimetry (53.8% versus 0%, P = .0005).

In another study of 78 chronic hypertensive patients, patients had uterine artery Dopplers performed at 24 weeks' gestation. Frusca et al¹⁸⁰ found that the presence of bilateral diastolic notches in the uterine arteries was significantly more prevalent among patients who developed superimposed preeclampsia (23% versus 0%, P <.001), or delivered SGA babies (85% versus 3%, P <.001). Of interest was the low prevalence of superimposed preeclampsia, which was only 3.8%. With regards to this issue, the authors commented that, although this may be due to the small number of patients, it could also reflect a benefit of low-dose aspirin therapy (50 mg/day), which was prescribed to all patients from the 12th week of gestation until delivery.

Risk of Stillbirth

Smith et al recently performed a very large, screening study that examined maternal characteristics and uterine artery Doppler findings in over 30,000 women.¹⁸¹ They sought to relate the risk of antepartum stillbirth to uterine artery Doppler flow velocimetry at 22 to 24 weeks. There were 109 stillbirths, and these were divided into placental (due to abruption, preeclampsia, or growth restriction) or unexplained causes. The risk of placental stillbirth was increased among women with a mean PI in the top decile (adjusted hazard ratio 5.5, 95% CI = 2.8 to 10.6) and those with a bilateral notch (adjusted hazard ratio 3.9, 95% CI = 2.0 to 7.8). Placental stillbirths occurred at earlier gestational ages than unexplained stillbirths: median (interquartile range); 30 (26 to 36) versus 38 (36 to 40) weeks, P <.001. Abnormal uterine artery Doppler was a better predictor of the risk of stillbirth due to placental causes (than unexplained stillbirths). Consequently, abnormal uterine artery Doppler was a good predictor of stillbirth at extreme preterm gestations, but a poor predictor of stillbirth at term.

Doppler of the uterine arteries has also been evaluated as a test to identify, among pregnancies already complicated by preeclampsia, those with a higher maternal and fetal risk. Van Asselt et al¹⁸² evaluated 28 women with severe preeclampsia and 80 with mild preeclampsia. Severe preeclampsia was defined as systolic blood pressure greater than or equal to 160 mm Hg and diastolic greater than or equal to 110 mm Hg, and/or proteinuria of +3 (≥3.0 g/L). Mild preeclampsia was defined as systolic blood pressure equal to 145 to 155 mm Hg and diastolic equal to 90 to 105 mm Hg, and proteinuria +1 or +2 (>0.3 and <3.0 g/L). Uterine artery Doppler waveforms were obtained with color Doppler imaging, and a mean PI of both uterine arteries above 1.20 was considered abnormal. Umbilical artery flow velocity waveforms were also obtained during the examination, where abnormal was defined as a PI greater than 2 SD above the mean for gestational age. The results of the last examination before delivery were correlated to perinatal outcome, mean maternal blood pressure, and degree of proteinuria.

The authors found that the uterine artery PI was positively correlated with the degree of proteinuria.¹⁸² Patients with abnormal uterine and umbilical artery waveforms were more likely to deliver prematurely: (1) 38.6 ± 2.5 weeks for normal uterine and umbilical artery waveforms; (2) 36.5 ± 2.5 weeks for abnormal uterine and normal umbilical artery waveforms; (3) 36.6 ± 3.3 weeks for normal uterine and abnormal umbilical artery waveforms; and (4) 31.7 ± 5.1 weeks for abnormal uterine and umbilical artery waveforms; P = .0005). Patients with abnormal uterine and umbilical artery waveforms also delivered more SGA fetuses: (1) 5.3% for normal uterine and umbilical artery waveforms; (2) 33.3% for abnormal uterine and normal umbilical artery waveforms; (3) 25% for normal uterine and abnormal umbilical artery waveforms; and (4) 77.8% for abnormal uterine and umbilical artery waveforms; P = .0001). Finally, patients with abnormal uterine and umbilical artery waveforms also had neonates with more intensive care unit admissions: (1) 15.8% for normal uterine and umbilical artery waveforms; (2) 33.3% for abnormal uterine and normal umbilical artery waveforms; (3) 62.5% for normal uterine and abnormal umbilical artery waveforms; and (4) 88.9% for abnormal uterine and umbilical artery waveforms; P = .0001). Women with a unilateral notch in the uterine arteries were 8 times more likely to deliver an SGA newborn (37% versus 6.5%; OR = 8.5, 95% CI = 2.7 to 27), and 10 times more likely to do so when bilateral notches were present (42% versus 6.5%; OR = 10.6, 95% CI = 3.2 to 35.2). When bilateral notches were present, there was also a 4-fold increased risk for a cesarean section (68.4% versus 33.9%; OR = 4.2, 95% CI = 1.5 to 12.2), and admission to the neonatal intensive care unit (52.6% versus 19.4%; OR = 4.6, 95% CI = 1.6 to 13.3).

Similar findings were reported by Joern and Rath,¹⁸³ who examined 142 pregnancies complicated by preeclampsia/ HELLP syndrome and/or SGA, and by Hofstaetter et al,¹⁸⁴ who studied a heterogeneous group of 421 high-risk pregnancies (SGA, preeclampsia, PIH, diabetes mellitus, prolonged pregnancy, third trimester hemorrhage, premature rupture of membranes, decreased fetal movements, poor obstetric history, poly- or oligohydramnios, fetal heart arrhythmia, and abnormal cardiotocogram). When both uterine and umbilical arteries were abnormal, the outcome was poorer, with a higher rate of preterm delivery and lower birth weight. However, outcome was not significantly affected when only the umbilical arteries were pathologic.

Likewise, in a cohort study of 186 cases of preterm preeclampsia diagnosed on average at 31.3 ± 3.6 weeks and delivered at 32.8 ± 3.3 weeks, those with abnormal uterine artery Doppler findings on admission had significantly lower gestational age at delivery (32.5 versus 35.3 weeks), and higher rates of very low birth weight (52% versus 5%) and of fetal growth restriction (70% versus 23%).¹⁸⁵

Another recent area of application is the prognostic value of uterine artery Doppler velocimetry in patients with SGA fetuses. The finding of an abnormal Doppler pattern strongly correlates with an adverse maternal and/or perinatal outcome.

Vergani et al investigated the use of uterine artery Doppler waveform analysis in growth-restricted fetuses (abdominal circumference <10th percentile) delivered at greater than or equal to 34 weeks.¹⁸⁶ Neonatal outcomes were compared in cases with normal versus abnormal Dopplers of the uterine arteries (mean RI >0.58 and/or presence of bilateral notching). Doppler of the uterine arteries was performed at the diagnosis of fetal growth restriction, and results were not used for patient management. The study showed that neonates (in cases of abnormal uterine Dopplers): (1) were more frequently born by cesarean section, particularly for nonreassuring fetal testing (27% versus 10%, P <.001); (2) had significantly lower gestational age at delivery (37.7 ± 2.0 versus 38.8 ± 1.6 weeks, P <.001); and (3) had lower birth weight percentiles (4.8 ± 5.1 versus 9.3 ± 10.2, P <.001). More importantly, they had a significantly greater risk of admission to the intensive care unit for reasons other than low birth weight alone (36% versus 11%, OR = 3.2, 95% CI = 1.9 to 5.3).

Several randomized controlled trials have assessed the value of screening with the uterine and umbilical arteries. Newnham et al¹⁸⁷ randomized women in the third trimester, and therefore did not assess the effect of screening early in pregnancy. In addition, the development of preeclampsia and SGA were not evaluated as outcomes in that study. Davies et al¹⁸⁸ randomized 2600 unselected women to Doppler (umbilical and uterine) or control groups at 19 to 22 weeks of gestation. The control group did not receive Doppler sonography. Women in the Doppler group had the first Doppler scan at 19 to 22 weeks; if abnormal values were obtained for the uterine circulation, the test was repeated at 24 weeks before being deemed abnormal. In the Doppler group, low-risk women had Dopplers performed once more at 32 weeks, while high-risk women (defined as those with hypertension, diabetes, previous SGA live birth, previous stillbirth or neonatal death, hypertension in a previous pregnancy or at booking, smoking) had Doppler examinations every month. Hypertension developed in fewer women in the Doppler group (8% versus 11%; RR = 0.70, 95% CI = 0.54 to 0.89). No improvement in perinatal outcome was demonstrated in pregnancies allocated to the Doppler group. The authors speculated that, although the reduction of hypertension in the Doppler group could have been the result of an action taken in response to the abnormal uterine artery examinations in the second trimester, it could have also been a chance finding, because there was no defined intervention instituted.

The combination of uterine artery Doppler velocimetry with other methods to evaluate maternal adaptation to pregnancy has been proposed to overcome the low PPV of the test.¹⁸⁹ Other tests include 24-hour automated maternal blood pressure monitoring¹⁹⁰ or biochemical markers such as α-fetoprotein,^{191,192, and 193} β-human chorionic gonadotropin (β-hCG),¹⁹⁴ a combination of maternal serum α-fetoprotein, β-hCG, and uric acid levels,¹⁹⁵ the N-terminal peptide of proatrial natriuretic peptide,¹⁹⁶ and urinary kallikrein:creatinine ratios.¹⁹⁷ Most of the studies that have examined uterine artery Doppler in combination with maternal serum markers have concentrated on the second trimester, although later we will discuss first trimester biochemical serum markers as well.

Uterine Artery Doppler Velocimetry Combined With 24-Hour Automated Blood Pressure Monitoring

Benedetto et al¹⁹⁰ conducted a case-control study to evaluate the value of adding 24-hour automated blood pressure (BP) monitoring to uterine artery Doppler screening in detecting a combination of PIH, preeclampsia, and SGA. Ninety women with singleton pregnancies at high risk for PIH, preeclampsia, or SGA, and abnormal uterine artery Doppler waveforms at 20 to 22 weeks (mean RI >0.58, diastolic notches) were compared with a control group (n = 90) with the same risk factors but normal uterine artery Dopplers. All patients underwent 24-hour BP monitoring with a portable automated device immediately after recruitment. The primary end points of the study were the development of PIH or preeclampsia, and SGA. The study found that the overall incidence of PIH or preeclampsia with or without SGA, or of SGA alone, was significantly higher in the abnormal Doppler group than in the control group (42% versus 17%, P <.05). The highest incidence of PIH and preeclampsia, with or without SGA, occurred in the group with abnormal uterine Doppler associated with a systolic midline estimating statistic of rhythm (MESOR) of at least 111 mm Hg, or a diastolic MESOR of at least 68 mm Hg (51.1% versus 2.2%, P <.0001). This difference, although less marked, was also observed in the normal uterine Doppler group (31.6% versus 4.2%, P <.01). Abnormal uterine artery Doppler alone predicted PIH/preeclampsia with or without SGA, with a 73% sensitivity, 55% specificity, 27% PPV, and 90% NPV. However, the predictive value of the test improved 2-fold with the addition of a diastolic MESOR greater than or equal to 0.68: 64% sensitivity, 91% specificity, 62% PPV, and 92% NPV. Thus, the results suggest that screening pregnant women with a two-stage test (uterine artery Doppler plus 24-hour BP monitoring) may significantly increase the performance of uterine artery Doppler screening in identifying pregnancies at high risk for PIH and preeclampsia.

Uterine Artery Doppler Velocimetry in Patients With Elevated Maternal Serum α-Fetoprotein

In addition to the well-established association between elevated levels of maternal serum α-fetoprotein (MSAFP) and neural tube defects, this marker has also been linked to an increased risk of perinatal death, preterm delivery, and SGA. Three studies^{191,192, and 193} have shown that fetuses with idiopathic elevation of MSAFP above 2.5 multiples of the median (MOM) and abnormal uterine artery Doppler velocimetry are at increased risk for abnormal perinatal outcome, including PIH, SGA, and preterm birth.

In the first study, uterine artery Doppler velocimetry was performed at 18 to 22 weeks, and again at 24 to 26 weeks in a group of 98 pregnancies with elevated MSAFP and normal fetal anatomy by ultrasound.¹⁹¹ Fourteen percent of the fetuses died in the perinatal period, 21.4% were delivered prematurely, and 17.3% were SGA. A persistent diastolic notch after 24 to 26 weeks predicted 79% of the perinatal deaths. These findings were confirmed in the study by Bromley et al,¹⁹² who reported poor outcome in 47% of 199 second trimester pregnancies with elevated MSAFP levels and severe early uterine artery diastolic notch at 18 to 19 weeks. Similarly Konchak et al,¹⁹³ in a study of 103 patients with unexplained elevated MSAFP levels, found that abnormal uterine artery Doppler velocimetry performed at 17 to 22 weeks was associated with an increased risk of PIH, low birth weight, and preterm delivery.

Uterine Artery Doppler in Pregnancies With Elevated Free β-hCG

In a retrospective study of 329 women with elevated maternal serum-free β-hCG levels found during Down's syndrome screening, Palacio et al¹⁹⁴ reported that values greater than or equal to 5.0 MOM were associated with a high incidence of PIH, SGA, or intrauterine fetal death (RR = 2.3, 95% CI = 1.3 to 3.9). Twenty-six women were selected for a prospective study based on a free β-hCG value greater than or equal to 5.0 MOM at 15 to 18 weeks and a normal fetal karyotype. Patients had uterine and umbilical artery Doppler velocimetry performed at 20 weeks and every month thereafter. Abnormal outcome in this group was observed in 8 fetuses (30.8%), and included isolated fetal growth restriction, PIH (2 with associated SGA), and intrauterine fetal demise with SGA. The presence of a diastolic notch at 24 weeks predicted adverse perinatal outcome with a sensitivity of 88%, specificity of 94%, PPV of 88%, and NPV of 94%.

Combination of MSAFP, Maternal Serum hCG, and Uric Acid Levels

Jauniaux et al¹⁹⁵ studied 41 singleton pregnancies that were selected because of bilateral diastolic notches in the flow velocity waveforms and/or an increased PI above the 95th centile between 20 and 24 weeks of gestation. MSAFP, maternal serum hCG (MShCG), and uric acid levels were measured and results compared with pregnancy outcome. All women were followed monthly until delivery. Among 20 patients with normal outcome, 19 had normal uterine artery Doppler velocimetry at their second visit. Twenty-one pregnancies had abnormal outcomes: severe proteinuric PIH with SGA (n = 8), isolated SGA (n = 8), mild PIH with normal fetal growth (n = 3), and placental abruption (n = 2). MSAFP, MShCG, and uric acid levels were increased in the first visit among pregnancies with adverse perinatal outcomes (MSAFP: 2.9 ± 0.8 versus 1.7 ± 0.9 MOM, P <.01; MShCG: 2.8 ± 1.1 versus 1.4 ± 0.9 MOM, P <.05; uric acid: 3.9 ± 1.1 versus 2.9 ± 0.7, P <.01). MSAFP and MShCG dropped by the second visit, whereas uric acid levels kept increasing in complicated pregnancies until delivery. Sixty-four percent (7 of 11) of the patients with PIH presented with a combination of high MSAFP and MShCG at the first visit; no patients with normal outcome presented with elevation of both markers (P <.01).

In contrast, Audibert¹³⁰ assessed the performance of prediction for preeclampsia by combining second trimester hCG, AFP, and uterine Doppler velocimetry in 2615 women. Overall, the different combinations of these tests showed poor sensitivities (6% to 8%), compared with those of the tests alone (16% to 39%). However, the specificities of the different combinations of tests (approximately 99.5%) were higher than those of the tests alone (88% to 95%).

Combination of Inhibin A and Uterine Artery Doppler

Inhibin A is a member of the transforming growth factor β family, and is predominantly secreted by the placenta and localized within the syncytiotrophoblast in pregnancy.¹⁹⁸ Elevated concentrations of inhibin A have been observed in preeclamptic women, compared to normotensive controls.¹⁹⁹ It has been shown that combining the measurements of second trimester uterine artery Doppler and maternal serum inhibin A improves the screening efficacy for the prediction for preeclampsia.²⁰⁰ Aquilina et al examined 689 women by obtaining inhibin A levels (15 to 19 weeks' gestation) with subsequent Doppler of both uterine arteries (18 to 22 weeks' gestation). The uterine artery RI was measured, and the presence or absence of notches in the flow velocity waveform was noted. Using bilateral notches/mean RI greater than or equal to 0.65, the sensitivity and false-positive rate for preeclampsia was 60% and 7%, respectively, while the positive likelihood ratio was 8.6 (95% CI = 5.7 to 12.6). However, for the same false-positive rate, when bilateral notches/mean RI greater than or equal to 0.55 and unilateral notches/mean RI greater than or equal to 0.65 were combined with inhibin A greater than or equal to 1.0 MOM, the sensitivity for preeclampsia improved to 71% and the positive likelihood ratio to 10.8 (95% CI = 7.4 to 15.4). The improvement in sensitivity for the combined method (compared to either inhibin A or uterine artery Dopplers alone) was statistically significant for both preeclampsia and preterm preeclampsia.

Pregnancy-associated plasma protein A, Free β-hCG, Activin A, Inhibin A, and Uterine Artery Doppler

Pregnancy-associated plasma protein A (PAPP-A) is a large glycoprotein complex predominantly produced by the placenta that has been implicated to play a role in the autocrine and paracrine control of trophoblast invasion of the decidua.²⁰¹ Activin A is a homodimeric glycoprotein produced by the decidua, placenta, and fetal membranes,²⁰² which is involved in cellular differentiation, proliferation, remodeling, and morphogenesis.²⁰³ In 1995, it was reported that maternal serum activin A concentration was abnormally high in patients with preeclampsia.²⁰⁴

Spencer et al assessed uterine artery Doppler measurements (mean PI), and serum levels of PAPP-A, free β-hCG, activin A, and inhibin A at 22.0 to 24.6 weeks' gestation.²⁰⁵ Compared to 144 women with normal pregnancy outcome, all 4 markers were significantly increased in the 24 women with subsequent preeclampsia. For an FPR of 5%, the detection rate for preeclampsia was 50% for uterine artery Doppler measurements, 5% for PAPP-A, 10% for free β-hCG, 44% for activin A, and 35% for inhibin A. However, by combining uterine artery mean PI with activin A and inhibin A, the detection rate for preeclampsia was increased to 75% (FPR 5%). Therefore, screening for preeclampsia by uterine artery PI at 22.0 to 24.6 weeks gestation can be improved by measurement of activin A and inhibin A levels as well.

In contrast, Ay et al²⁰⁶ examined different combinations of fetal-placental proteins (AFP, hCG, estriol, inhibin A, activin A) and uterine artery Doppler in 178 healthy pregnant women in the second trimester to analyze the predictive power in screening for preeclampsia. Although the addition of either inhibin A or activin A slightly improved the predictive accuracy of uterine artery Doppler, this improvement was not considered to be important clinically.

Placental Growth Factor, Soluble Fms-Like Tyrosine Kinase, and Uterine Artery Dopper

There is growing evidence that an imbalance between angiogenic factors, such as vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), and factors that inhibit angiogenesis (such as soluble Fms-like tyrosine kinase [sFlt-1]) are closely related to the pathogenesis of preeclampsia.²⁰⁷ Preeclampsia has been proposed to be an antiangiogenic state. The angiogenic factors are involved in the regulation of placental vascular development and maternal endothelial function by binding to Flt-1. An alternative splicing of Flt-1 pre-mRNA produces sFlt-1, which is a soluble, circulating version of the protein and is a potent antagonist of PlGF and VEGF.²⁰⁷ Studies have shown that a marked increase in sFlt-1 plasma concentration is observable weeks before the clinical manifestation of preeclampsia and IUGR (intrauterine growth restriction), while conversely, PlGF plasma concentration is lower in women in whom preeclampsia will occur.²⁰⁸

Our group recently conducted a large, prospective cohort study in a low-risk population to determine the role of the combined use of uterine artery Doppler velocimetry and maternal plasma PlGF and sVEGFR-1 (soluble vascular endothelial growth factor receptor-1, also known as sFlt-1) concentrations in the second trimester for the identification of patients at risk for severe and/or early onset preeclampsia (≤34 weeks' gestation).²⁰⁹ A total of 3348 pregnant women were studied, and plasma samples were obtained between 22 and 26 weeks' gestation at the time of ultrasound examination. Abnormal uterine artery Doppler velocimetry was defined as the presence of bilateral notches and/or a mean PI above 95th percentile for gestational age. Abnormal uterine artery Doppler and a maternal plasma PlGF of less than 280 pg/mL were independent risk factors for the occurrence of preeclampsia, severe preeclampsia, and early onset preeclampsia. In addition, among patients with abnormal uterine artery Doppler velocimetry, PlGF concentration contributed significantly in the identification of patients destined to develop early onset and/or severe preeclampsia. The combination of abnormal uterine artery Doppler and PlGF of less than 280 pg/mL was associated with an OR of 8.6 (95% CI = 5.35 to 13.74) for the development of preeclampsia, an OR of 43.8 (95% CI = 18.48 to 103.89) for the development of early onset preeclampsia, and an OR of 37.4 (95% CI = 17.64 to 79.07) for the development of severe preeclampsia. It was established that the sequential determination of uterine artery Doppler velocimetry and maternal plasma concentrations of PlGF, or vice versa, could identify women who are at high risk for early onset and/or severe preeclampsia. In addition, in contingency screening, the predictive performance of the combined test could be achieved by offering the determination of PlGF concentrations only to those women with abnormal uterine artery Doppler results (about 10% of the population). In contrast, maternal plasma sVEGFR-1 concentration was of limited use in the prediction of early onset and/or severe preeclampsia.

Our group also conducted a cross-sectional study to determine sVEGFR-1 concentrations in normal pregnant women, women with SGA fetuses, and patients with preeclampsia.²¹⁰ Patients with SGA fetuses and those with preeclampsia were subclassified according to the results of uterine and umbilical artery Doppler velocimetry examinations. Mothers with SGA fetuses had an sVEGFR-1 concentration higher than normal pregnant women, but lower than patients with preeclampsia. However, among patients with SGA fetuses, only those with abnormal uterine artery Doppler velocimetry had a mean plasma sVEGFR-1 concentration significantly higher than normal pregnant women.

Diab et al conducted a study in 108 pregnant women with abnormal uterine artery Doppler velocimetry at 23 weeks, to determine whether measuring plasma levels of sFlt-1 and PlGF could improve the prediction of preeclampsia and IUGR.²⁰⁷ There were 33 cases of preeclampsia and 9 cases of IUGR. In women whose pregnancies became complicated by preeclampsia and/or IUGR, sFlt-1 levels were significantly higher (1337.2 ± 986.3 versus 513.3 ± 72.6 pg/mL, P <.001) and PlGF levels were significantly lower (119.5 ± 23.5 versus 172.8 ± 23.2 pg/mL, P <.001), as compared to normal pregnancies. Importantly, these alterations were more pronounced in cases of early onset preeclampsia and in cases of IUGR that necessitated delivery before 34 weeks, suggesting that defective angiogenesis may be especially important in such pregnancies. By using an sFlt-1/PlGF ratio, all of these complications (preeclampsia and IUGR) could be predicted with 98% sensitivity, 95% specificity, 93% PPV, and 98% NPV. Therefore, by measuring uterine perfusion, sFlt-1, and PlGF concurrently in the second trimester, this improved the prediction of preeclampsia and IUGR.

Homocysteine and Uterine Artery Doppler

In 406 low-risk women, Onalan et al studied the predictive efficacy for preeclampsia of a combination of second trimester maternal serum homocysteine and uterine artery Doppler velocimetry at 20 weeks' gestation.²¹¹ A homocysteine cutoff level of 6.3 μmol/L (95th centile) was used. The addition of hyperhomocysteinemia to Doppler alterations improved the predictive accuracy of individual tests. The sensitivities of homocysteine, uterine artery Doppler, and the combined method for preeclampsia were 56%, 44%, and 61%, respectively. The specificities of homocysteine, uterine artery Doppler, and the combined method for preeclampsia were 94%, 93%, and 98%, respectively. Corresponding positive and negative likelihood ratios were (10.3 and 0.36), (6.7 and 0.69), and (16.1 and 0.76), respectively.

In summary, it appears that based on the evidence, a combination of either placental proteins, or angiogenic factors, or homocysteine with uterine artery Doppler velocimetry slightly improves the predictive ability for preeclampsia compared to the tests alone. However, large, prospective studies are necessary before claims can be made that combinations of tests are effective to predict preeclampsia.

Uterine artery Doppler velocimetry, performed at admission for preterm labor, has been shown to correlate significantly with failed tocolysis, a short interval between admission and delivery, and abnormal perinatal outcome. Brar et al²¹² studied the uterine and umbilical artery waveforms of 60 patients in preterm labor, receiving either ritodrine (n = 20) or magnesium sulfate tocolysis (n = 40). A uterine artery S/D ratio greater than 2.6 or the presence of a diastolic notch was considered abnormal. Fifty-eight percent (7 of 12) of the patients with abnormal uterine artery Doppler velocimetry failed tocolysis and delivered within 48 hours of the examination, as compared to 14.6% (7 of 48) of the patients with a normal waveform (P <.01).

In a subsequent study, 92 patients had uterine and umbilical artery waveforms determined at admission.²¹³ The outcome variables were preterm delivery (before 37 weeks), SGA rate, rate of cesarean section for fetal distress, number of days in the neonatal intensive care unit, and neonatal death rate. An abnormal uterine artery Doppler waveform had a PPV of 79%, and an NPV of 69% for preterm delivery. The PPV and NPV for abnormal perinatal outcome were 64% and 82%, respectively.²¹⁴

These results were confirmed in a case-control study of 55 patients in preterm labor, who were compared with 30 control patients (not in labor).²¹⁵ No difference in uterine artery impedance to flow was observed between the study and control groups. However, 90% (10 of 11) of the patients with a pathologic uterine artery velocity waveform (PI >0.90) delivered preterm, whereas only 36% (16 of 44) of the patients with a normal PI did so (P <.05). The PPV and NPV for preterm delivery in this series were 90.9% and 63.6%, respectively.²¹⁵

The effect of different tocolytic agents on uteroplacental blood flow has also been investigated. Tocolysis with either β-adrenergic drugs or magnesium sulfate is associated with a reduction in the impedance to blood flow in the uteroplacental circulation.^213,216 In the case of β-adrenergic drugs, the reduction in impedance to flow may be caused by a reduction in peripheral resistance.²¹⁴ In the case of magnesium sulfate, it is likely that the decreased impedance to flow is a result of a reduction in peripheral uterine vasculature resistance.^216,217 However, short-term tocolysis with nifedipine does not appear to have an influence on either the fetal or the uteroplacental circulations when evaluated by Doppler velocimetry.²¹⁸

Rizzo et al²¹⁹ studied multiple vessels (uterine, umbilical, fetal thoracic descending aorta, renal artery, and middle cerebral artery) with Doppler velocimetry in patients with preterm labor. Patients in preterm labor had higher uterine RI values (when compared with normal reference values for gestational age), but no difference was found between those who delivered within 48 hours or after 48 hours of admission (0.75 ± 0.95 versus 0.85 ± 76, P = .681). However, PI values from the middle cerebral artery were significantly lower in the group of fetuses who delivered within 48 hours (1.80 ± 1.32 versus −0.39 ± 1.19, P ≤ .001). The authors concluded that preterm delivery is associated with alterations in fetal cerebral waveforms, and knowledge of these changes may prove useful in the evaluation of patients with preterm labor.

Hyperglycemia may impair placental development and hence predispose the pregnancy to adverse perinatal outcome. Several studies have assessed uteroplacental perfusion with Doppler.^{220,221,222,223,224,225, and 226}

It has been demonstrated that insulin-dependent diabetic pregnancies do not show the normal decline in uterine artery PIs during the third trimester.²²⁵ Absolute PI values, however, remain within normal limits and do not help predict perinatal outcome. In contrast, increased discordance in the uterine artery S/D ratios during the third trimester (≥0.6) may be of value in identifying patients at risk for fetal distress.²²⁶ A compilation of data from 5 other studies, including 217 diabetic patients (68 class A, 73 class B, 46 class C, 11 class D, 8 class F, and 11 class R, according to White's classification²²⁷) who had serial Doppler velocimetry studies of the uterine arteries performed throughout pregnancy, showed the test to be within normal limits, regardless of the diabetic class, except in diabetic pregnancies complicated by preeclampsia.

Uterine artery Doppler in the third trimester independently identifies, among pregnancies with diabetes and vasculopathy, those at increased risk of preeclampsia and greater risk of cesarean delivery for nonreassuring fetal status.²²⁸ The information provided by uterine artery Doppler operates along a continuum, so that in women with diabetes antedating pregnancy, the risk for adverse perinatal outcome (operative delivery for fetal distress, preterm delivery, 5-minute Apgar score less than 5, low umbilical artery pH, and SGA infants) is directly correlated with the severity of the abnormal uterine artery Doppler findings.²²⁹ In addition, among women with gestational diabetes, abnormal uterine artery Doppler findings are significantly associated with greater risk of subsequent preeclampsia (positive LR 10.4, 95% CI = 6.6 to 10.4).²³⁰

Association With Miscarriage

Approximately 20% of pregnant women have symptoms of threatened abortion in the first trimester of pregnancy, and approximately half of these will ultimately have a miscarriage. Due to the high prevalence of spontaneous abortion in this group, a technique that could predict pregnancy outcome would be extremely helpful. Accordingly, studies have been performed using Doppler of the uteroplacental circulation to differentiate which cases of threatened abortion would end in miscarriage; however, many have found that these were not useful predictors of pregnancy outcome.^{231,232, and 233}

One study found no difference in the RI of the uterine, arcuate, and radial arteries between pregnancies with threatened abortion and normal pregnancy outcome, and women with abnormal outcome.²³¹ Alcázar et al²³² performed a prospective study and compared 49 women with first trimester threatened abortion and a living embryo, to 129 women with normally developing first trimester pregnancies. Measurements of the peak systolic velocity and PI of the uterine arteries and spiral arteries were performed. There were no differences found in any Doppler parameter assessed between the study group and the controls, even in those pregnancies that ended in spontaneous abortion. Pellizzari et al²³³ assessed uterine artery blood flow (PI, RI, peak systolic velocity) in normal first trimester pregnancies and in those complicated by uterine bleeding. Of 46 patients affected with uterine bleeding, 18 had an incomplete miscarriage, 8 had blighted ovum, 5 had a missed miscarriage, and 15 delivered term, liveborn infants. There were no significant differences found in any of the 3 vascular indices between the normal and pathological groups of patients.

On the other hand, discordant PI values of the 2 uterine arteries (determined as early as 6 weeks' gestation) may be a predictor of subsequent miscarriage in early viable pregnancies, secondary to insufficient uteroplacental circulation.²³⁴ One study found that miscarriage was significantly associated with the magnitude of the difference in PI values between the 2 uterine arteries. However, the location of the placenta was not reported for the ongoing pregnancies or the ones that ended up in miscarriage. It was suggested that the difference in the impedance indices of the uterine arteries could be caused by the location of the trophoblastic reaction, for which one artery develops the physiologic changes before the contralateral artery. In conclusion, from the studies reviewed, it seems therefore that Doppler evaluation of the uterine circulation cannot reliably predict if, or when, a pregnancy will end in subsequent miscarriage.⁶²

With regard to missed abortions, Giacobbe et al²³⁵ reported on the difference in uteroplacental Doppler velocimetry between normal pregnancies and missed abortions (including anembryonic pregnancies), with findings similar to previous studies. They found significantly higher RI and PI values of the uterine arteries in normal pregnancies than in missed abortions.

Uterine Artery Doppler in the First Trimester as a Screen for Preeclampsia

As previously described, multiple studies have been performed to assess the validity of second trimester uterine artery Doppler examination as a screening tool for preeclampsia. However, the observation that trophoblast invasion is maximal in the first trimester and that preeclampsia derives from a relative failure of this event, justifies the evaluation of uterine artery Doppler findings in the first trimester of pregnancy as well.²³⁶ Moreover, the ability to predict women who are likely to develop preeclampsia before the onset of the disease is undoubtedly an important goal.¹¹⁴ This would not only identify those women who require closer antenatal surveillance, but would also help in selecting those most likely to benefit from any therapeutic measures.

First Trimester Uterine Artery Doppler

In the majority of normal pregnancies presenting with high uterine artery PI in the first trimester, there is normalization in impedance to flow with advancing gestation, presumably owing to progressive physiological trophoblastic invasion.⁷⁰ By contrast, however, in those pregnancies destined to develop preeclampsia owing to impaired trophoblastic invasion, the uterine artery PI is increased in both the first and second trimesters of pregnancy. Therefore, multiple studies have prospectively evaluated the role of first trimester uterine artery Doppler examination in large populations.^{176,236,237, and 238}

Martin et al performed uterine artery Doppler velocimetry at 11 to 14 weeks in 3045 pregnancies.²³⁷ Velocity waveforms were obtained of both arteries, and the mean PI of the 2 arteries was determined. The predictive value of a mean PI greater than the 95th centile in the prediction of preeclampsia and/or fetal growth restriction was calculated. There were 63 cases of preeclampsia, and 290 cases of fetal growth restriction. For an FPR of 5%, the sensitivity of mean PI greater than the 95th centile for preeclampsia (with or without fetal growth restriction) was 27%, but for fetal growth restriction alone it was 12%. Importantly, there was a stepwise increase in sensitivity with the severity of preeclampsia, as expressed by gestational age at delivery. Therefore, the respective sensitivities for these same complications (requiring delivery before 36 weeks' gestation) were higher (40% and 22%, respectively). Furthermore, the respective sensitivities for these same complications (requiring delivery before 32 weeks' gestation) were even higher still (60% and 28%, respectively). Therefore, the finding that the sensitivity for both preeclampsia and fetal growth restriction was inversely related to the gestational age at delivery demonstrates that uterine artery Doppler assessment is much better at identifying the more severe (and therefore most clinically relevant) cases of preeclampsia.

Recently in 2007, a large, prospective study from the Fetal Medicine Foundation was performed to screen for hypertensive disorders of pregnancy utilizing maternal characteristics and uterine artery Doppler PI at 11.0 to 13.6 weeks gestation.¹⁷⁶ Of 6015 patients, 107 developed preeclampsia and 107 developed gestational hypertension. When utilizing screening by first trimester uterine artery PI, the detection rate of preeclampsia was 41%, for a 10% FPR. However, by combining uterine artery Doppler with maternal factors, the detection rate was improved to 62%. The study found that screening by first trimester uterine artery PI alone was particularly effective in identifying women who developed severe early onset preeclampsia, rather than late onset disease, gestational hypertension, or SGA. For an FPR of 10%, the predicted detection rate of preeclampsia requiring delivery before 34 weeks was 82% (compared to 31% for late preeclampsia, 12% for gestational hypertension, and 18% for SGA).¹⁷⁶ This is especially important because it is early rather than late preeclampsia, which is associated with an increased risk of perinatal morbidity/mortality and both short-term and long-term maternal complications.

A recent, large, prospective study examined the relationship between uterine artery Doppler findings (RI and diastolic notch) at 11 to 14 weeks, and the development of both term and preterm preeclampsia.²³⁶ Of 3,058 pregnancies examined, there were 57 and 33 cases of term and preterm preeclampsia, respectively. When using cutoffs of RI greater than the 90th and greater than the 95th centile, the sensitivity of uterine Doppler for detecting preterm preeclampsia was 49% and 24%, respectively. Furthermore, there was a 6-fold increase in the likelihood ratio for developing preterm preeclampsia when the uterine Doppler RI was greater than the 90th centile. The RI was significantly higher in those women who subsequently developed preterm preeclampsia (mean RI, 0.79), as compared to those who developed term preeclampsia (mean RI, 0.72; P = .002) or who had a normal outcome (mean RI, 0.70; P = .0001). Also, the prevalence of first trimester bilateral uterine artery notches in preterm preeclampsia (76%) was significantly higher than in pregnancies with term preeclampsia (51%, P = .036) or with a normal outcome (45%, P = .0001). However, because of the high prevalence of this finding in normal pregnancies, the presence of bilateral notches was a relatively poor predictor of preeclampsia. This finding is consistent with prior studies, which concluded that early bilateral notching alone is unlikely to be useful in screening for pregnancy complications.^237,238 When comparing women with a normal outcome to those who developed term preeclampsia, there were no significant differences in first trimester mean uterine artery RI or prevalence of bilateral notches. The data suggested that preterm preeclampsia was strongly associated with defective invasion of the spiral arteries, in contrast to the findings in term preeclampsia, which may be a result of placental deterioration at term.²³⁶ This supports the hypothesis of a distinct etiology in early onset versus late onset preeclampsia. Despite their findings, the authors concluded that while there is a strong relationship between first trimester uterine artery Doppler indices and the subsequent development of preterm preeclampsia, their data did not support its routine introduction into clinical practice.

Gómez et al prospectively examined first trimester (11 to 14 weeks) uterine artery Doppler indices in 999 low-risk women.²³⁸ The mean PI and the presence of bilateral diastolic notching were recorded. There were 22 cases of preeclampsia and 37 cases with intrauterine growth restriction (birth weight less than the 5th percentile). By using the 95th centile in mean uterine artery PI as a cutoff, there was a 24% sensitivity for preeclampsia, and 24% sensitivity for fetal growth restriction. A population with mean uterine artery PI less than the 95th percentile had a 5-fold increased risk of subsequently developing hypertensive disorders and fetal growth restriction. The authors concluded, based on their results, that the use of a single uterine artery Doppler measurement for screening purposes in unselected early pregnancy populations had limited clinical value.

In response to the paper by Gómez et al, Stuart Campbell wrote an opinion in 2005 (titled "First Trimester Screening for Preeclampsia"), where he makes the argument against this statement.²³⁹ First, he points out that severe preeclampsia occurs in about 1 in 200 gestations, and very severe preeclampsia in about 1 in 300 gestations. Therefore, this is a high enough prevalence to justify antenatal screening. Second, transabdominal examination of the uterine artery in the first trimester is acceptable, simple, and safe. Third, he asks the question of whether the performance of uterine artery Doppler first trimester screening is good enough, and discusses the large study (>3000 pregnancies) by Martin et al,²³⁷ which is described above. He states that the technique is effective at defining a population of women who will develop severe/very severe preeclampsia/fetal growth restriction. Campbell also makes the argument that the false-positive rate of 5% may not be the best cutoff for screening for preeclampsia. This is because in the case of preeclampsia, the aim of screening is to help stratify the intensity of subsequent antenatal surveillance. Altough there can be anxiety associated with being "classified" as being at high risk for preeclampsia, such a classification does not carry a direct risk to the mother or fetus. Therefore, in screening for preeclampsia, it would be reasonable to maximize the detection rate, at a cost of a higher screen-positive rate. This is in contrast to Down's syndrome screening programs where the FPR has to be kept low, because a positive result can lead to an invasive procedure with the potential risk for miscarriage. Fourth, he discusses that uterine artery Doppler screening can be "added on" to an existing test already being performed (ie, standard ultrasound screening), which does not involve significant costs, and will identify a cohort of at-risk patients who would benefit from increased surveillance. Finally, Campbell points out that whatever effects treatment or increased surveillance may have in preventing or delaying preeclampsia/fetal growth restriction, they are more likely to be effective when started in the first trimester, before the secondary wave of trophoblast invasion has begun. Therefore, all these reasons justify the importance of early uterine artery Doppler screening.

Changes in Uterine Artery Doppler From First to Second Trimesters

A recent study by Plasencia et al evaluated the performance of screening for preeclampsia by uterine artery PI at 11.0 to 13.6 weeks' gestation, and the change in uterine artery PI (ratio) between 11.0 to 13.6 and 21.0 to 24.6 weeks' gestation.⁷⁰ They examined 3107 singleton pregnancies, of which 3% (93 of 3107) of the patients developed preeclampsia. Early preeclampsia was defined when delivery occurred prior to 34 weeks, whereas late preeclampsia was defined when delivery occurred at 34 weeks or more. At 11.0 to 13.6 weeks, the uterine artery PI was above the 90th centile in 77% of the early preeclampsia cases, and in 27% of the late preeclampsia cases. At 21.0 to 24.6 weeks, there was persistence of uterine artery PI above the 90th centile in 94% of the early preeclampsia cases, 74% of the late preeclampsia cases, and 37% of those who did not develop preeclampsia. The study demonstrated that along with maternal variables, both uterine artery PI (11.0 to 13.6 weeks) and the change in uterine artery PI (between 11.0 to 13.6 and 21.0 to 24.6 weeks) provided significant independent contributions to the prediction of preeclampsia. When utilizing screening by maternal variables and first trimester uterine artery PI, the detection rate of early preeclampsia was 45.5% (FPR 5%). However, by including the PI ratio, the detection rate was improved to 90.9%. Therefore, the predicted detection rates of early and late preeclampsia (using a combined screening of maternal variables, first trimester uterine artery PI, and change in uterine artery PI or ratio) were 90.9% and 31%, respectively (FPR 5%). Importantly, the same screening performance was achieved by reserving second trimester screening for only the 20% of women at the highest risk after first trimester screening (ie, contingency screening). They also noted that in pregnancies with a normal outcome, the decrease in uterine artery PI (between 11.0 to 13.6 and 21.0 to 24.6 weeks) was steeper, as compared to those developing preeclampsia. Because of their findings, Plasencia et al suggest that these results could provide the scientific basis for the rationalization of antenatal care.⁷⁰ Combined screening for preeclampsia by maternal factors and uterine artery Doppler could be undertaken in all women at their 11.0- to 13.6-week hospital visit. In the 20% of women with the highest risk for preeclampsia, Doppler assessment of the uterine arteries could be repeated in the second trimester. On the basis of findings in the first trimester and the change in uterine artery PI between the first and second trimesters, 75% (of the 20% of cases with the highest risk at 11.0 to 13.6 weeks) could be assigned to the low-risk group, and 25% (or 5% of the total) would remain in the high-risk group with the need for increased surveillance.

Gómez et al assessed the uterine arteries by Doppler sonography for the presence of bilateral notches at 11 to 14 weeks and 19 to 22 weeks in 870 pregnancies (including 27 that subsequently developed preeclampsia or gestational hypertension).²² At 11 to 14 weeks, bilateral notches were observed in 42% of cases, including 66.7% (18 of 27) that developed hypertensive disorders. At 19 to 22 weeks, there was persistence of bilateral notches in 61% (11 of 18) who developed hypertensive disorders, and in 21% (72 of 350) with no such disorders.

Similarly, Carbillon et al assessed the uterine arteries by Doppler sonography for the presence of either unilateral or bilateral notches at 12 to 14 weeks and 22 to 24 weeks in 243 pregnancies (including 12 that subsequently developed preeclampsia).²⁴⁰ Notches at 12 to 14 weeks were observed in 57% of cases, including 83% (10 of 12) that developed preeclampsia. At 22 to 24 weeks, there was persistence of notches in 80% (8 of 10) that developed preeclampsia, and in only 21% (27 of 129) that did not develop preeclampsia.

Uterine Artery Doppler in the First Trimester as a Screen for Fetal Growth Restriction

Several studies have reported on the sensitivity of first trimester uterine artery Doppler for the prediction of SGA.^237,238,241 By using the 95th centile as a cutoff for the mean uterine artery PI, 24% of SGA fetuses (≤5th centile) were identified.²³⁸

Similarly, Dugoff et al collected uterine artery Doppler velocimetry data (mean RI and diastolic notches) performed between 10 and 14 weeks gestation on 1008 women enrolled in the FASTER (First and Second Trimester Evaluation of Risk) trial.²⁴¹ IUGR was defined as birth weight below the 10th centile for gestation. They found that the mean uterine artery RI was equal on both sides (0.59 ± 0.14). Women having a high RI (≥75th percentile) were 5.5 times more likely to have IUGR (95% CI = 1.6 to 18.7). There was no significant relationship found between notching and IUGR. The sensitivity of RI greater than or equal to the 90th centile was 33% (specificity 89.8%), whereas for RI greater than or equal to the 95th centile the respective values were 16.7% and 95%. Therefore, elevated first trimester uterine artery mean RI was significantly associated with developing IUGR later in pregnancy. However, IUGR was only seen in 12 (1.2%) cases, due to the low-risk population enrolled, and therefore, this is an important limitation of the study. The authors speculate that first trimester uterine artery Doppler screening may identify those patients who may be at higher risk, and should then be followed up with a second trimester scan.²⁴¹

As discussed previously, Martin et al performed uterine artery Doppler velocimetry at 11 to 14 weeks in 3045 pregnancies.²³⁷ Velocity waveforms were obtained of both arteries, and the mean PI of the 2 arteries was determined. The predictive value of a mean PI greater than the 95th centile in the prediction of preeclampsia and/or fetal growth restriction was calculated. There were 290 cases of fetal growth restriction. The sensitivity of mean PI greater than the 95th centile for fetal growth restriction alone was 12%. However, the sensitivity for fetal growth restriction (requiring delivery before 32 weeks' gestation) was higher, at 28%.

Uterine Artery Doppler and First Trimester Biochemical Serum Markers

An increasing number of biochemical agents have been evaluated as first trimester markers for predicting preeclampsia. Preeclampsia is a multifactorial disease that is able to manifest itself in various pathophysiological aspects. Thus, the utilization of different biochemical markers seems sensible in light of the characteristic nature of preeclampsia.²⁴¹ These markers allow screening at a relatively early stage and, most importantly, show relatively high predictive values and improved diagnostic performance if combined with first trimester Doppler sonography.²⁴²

PAPP-A and First Trimester Uterine Artery Doppler

PAPP-A is involved in placenta growth and development. Maternal serum concentrations of PAPP-A have been shown to be relatively low in the first trimester of pregnancies complicated by SGA and/or preeclampsia, although the sensitivity of PAPP-A in screening for these conditions remains disappointingly low.²⁴³ Therefore, Pilalis et al examined the value of combining uterine artery Doppler velocimetry and measurement of maternal serum PAPP-A at 11 to 14 weeks in the prediction of preeclampsia and/or SGA.²⁴³ There were 878 women prospectively studied. Mean uterine artery PI greater than or equal to the 95th centile and PAPP-A less than or equal to the 10th centile each predicted 23% of the women that developed preeclampsia. For SGA less than or equal to the 5th centile, mean uterine artery PI greater than or equal to the 95th centile predicted 23% of cases, while PAPP-A less than or equal to the 10th centile predicted 34%. By combining uterine artery Doppler and PAPP-A, the prediction of SGA less than or equal to the 5th centile was better than uterine artery Doppler alone; however, the difference was not statistically significant (P = .07). Although increased uterine artery PI was found to be an independent risk factor for preeclampsia (OR 2.76, 95% CI = 1.11 to 6.81, P = .02), this was not the case for PAPP-A. Therefore, the authors suggest that the role of PAPP-A in the prediction of preeclampsia is limited.²⁴³

Combination of PAPP-A and Free β-hCG With Second Trimester Uterine Artery Doppler

An association exists between low first trimester maternal serum free β-hCG and PAPP-A, and subsequent development of pregnancy complications, such as preeclampsia, fetal growth restriction, and preterm delivery.²⁴⁴ Spencer et al measured maternal serum PAPP-A and free β-hCG at 11.0 to 13.6 weeks, and uterine artery PI at 22 to 24 weeks, in 4390 women.²⁴⁴ In the pregnancies that resulted in preeclampsia (64 cases), the median PAPP-A was lower and the median uterine artery mean PI was higher, but the median free β-hCG was not significantly different than in the normal outcome group. In screening for preeclampsia, for a 5% FPR, the detection rate was 14% for low PAPP-A, 55% for uterine artery mean PI greater than or equal to the 95th centile, and 62% for a combination of PAPP-A and uterine artery mean PI. Therefore, the combination of first trimester PAPP-A and uterine artery mean PI at 22 to 24 weeks improved the screening efficacy for the prediction of preeclampsia. However, the sensitivity of the 2 tests combined was not statistically different from that of uterine artery Doppler alone.

Combination of Placental Protein 13 With First Trimester Uterine Artery Doppler

Recently, placental protein 13 (PP-13) has been assessed as a first trimester marker of predicting preeclampsia. PP-13 is a member of the galectin family predominantly expressed by the placenta, specifically by the syncytiotrophoblast. PP-13 binds to proteins on the extracellular matrix between the placenta and the endometrium, and is thought to be involved in placental implantation and maternal artery remodeling.²⁴⁵ Preliminary data suggest that the gene for PP-13 is down-regulated in women with preeclampsia requiring early delivery.²⁴⁵

Accordingly, Nicolaides et al recently prospectively investigated the value of maternal serum PP-13 measurement and uterine artery Doppler PI during first trimester screening (11.0 to 13.6 weeks) in the prediction of early preeclampsia through a case-control study.²⁴⁶ There were 10 women who went on to develop early preeclampsia (requiring delivery before 34 weeks) and 423 unaffected controls. The median uterine artery PI was higher (1.43 MOM) and the median serum PP-13 level was lower (0.07 MOM) in the cases that developed preeclampsia (requiring delivery before 34 weeks), versus unaffected pregnancies. For a 10% FPR, the detection rate of preeclampsia requiring early delivery was about 80% for serum PP-13, 40% for uterine artery PI, and 90% for the combined markers. Importantly, the same high detection rate of 90% could be achieved by utilizing a policy of contingency screening (all women are primarily screened by PP-13, and only the 14% at highest risk are then screened by uterine artery Doppler).

First Trimester PAPP-A, PP-13, and Second Trimester Uterine Artery Doppler

In a case-control study of preeclampsia cases and matched controls, patients had PAPP-A and PP-13 measured from 11.0 to 13.6 weeks' gestation, and then underwent uterine artery Doppler flow velocimetry to measure the mean PI at 22 to 24 weeks' gestation.²⁴⁷ This is the first study to have measured in parallel PAPP-A and PP-13 in the first trimester as potential serum markers for the prediction of preeclampsia. There were 44 cases with early preeclampsia where delivery was induced prior to 35 weeks, and another 44 cases with preeclampsia in which delivery was not induced before term. In pregnancies with preeclampsia, PAPP-A (0.89 MOM) and PP-13 (0.69 MoM) were reduced, and with more severe or early preeclampsia the levels were further reduced. Accordingly, these markers may serve as useful tools for identifying subtypes of preeclampsia, and also enable the refinement of follow-up strategies of surveillance and treatment.²⁴⁷ For an FPR of 20%, PP-13 levels may be useful in predicting preeclampsia (sensitivity 44%) and early preeclampsia (sensitivity 50%), and the accuracy of the method increases when coupled with second trimester uterine artery Doppler PI measurements (sensitivities 74% and 79%, respectively). Although PAPP-A provides some prediction for preeclampsia when combined with uterine artery PI (sensitivity 76% for all cases of preeclampsia and 76% for early preeclampsia), it does not add to the prediction of early preeclampsia when PP-13 and uterine artery PI are used together (sensitivity 70%). This study yielded a lower prediction of early preeclampsia by PP-13 as compared to the study of Nicolaides et al (sensitivity 80%, FPR 10%).²⁴⁶ This variation in the detection rate was felt to be due to variation in ethnicity and parity between the 2 studies. Therefore, Spencer et al concluded that further studies would be required to establish the real value of PP-13 in first trimester screening for preeclampsia.²⁴⁷

Placental Growth Factor and First Trimester Uterine Artery Doppler

Placental growth factor (PlGF) is a proangiogenic protein, and preeclampsia is associated with reduced production of PlGF. These decreased levels are reported not only during the clinical phase of the disease, but they also precede the clinical onset of the disease and are evident from both the second and first trimesters of pregnancy.²⁴⁸ Therefore, a recent case-control study was performed to investigate the potential value of maternal serum PlGF in first trimester (11.0 to 13.6 weeks) screening for preeclampsia, and also to estimate the potential performance of screening for preeclampsia by a combination of maternal factors, uterine artery PI, and maternal serum PlGF. There were 4 outcome groups studied: 29 cases of early preeclampsia (required delivery before 34 weeks), 98 cases of late preeclampsia, 88 cases of gestational hypertension, and 609 normal controls. In the control group, the median PlGF was 0.991 MoM. In the early and late preeclampsia groups, PlGF was reduced (0.611 and 0.822 MOM, respectively), but in gestational hypertension (0.966 MOM), there were no significant differences from controls. By using maternal characteristics and obstetrical history, and PlGF, detection rates for early and late preeclampsia were 62% and 52%, respectively (FPR 10%). However, by using combined screening (maternal characteristics and obstetrical history, uterine artery PI, and PlGF), detection rates for early and late preeclampsia were 90% and 49%, respectively (FPR 10%).

There is an abundance of literature regarding Doppler velocimetry of the uterine arteries as a screening test to identify women at high risk for the development of preeclampsia and SGA fetuses. This is certainly appropriate, because uterine artery Doppler velocimetry meets some of the criteria for the ideal predictive test (simple, rapid, inexpensive, innocuous). For the most part, the most common screening strategy has been to scan patients in the second trimester. However, the observation that trophoblast invasion is maximal in the first trimester, and that preeclampsia derives from a relative failure of this event, justifies the evaluation of uterine artery Doppler findings in the first trimester of pregnancy as well. Moreover, the ability to predict as early as possible women who are likely to develop preeclampsia before the onset of the disease is undoubtedly an important goal. This would not only identify those women who require closer antenatal surveillance, but would also help in selecting those most likely to benefit from any therapeutic measures. Accordingly, it is now clear that first trimester screening using uterine artery Doppler velocimetry (alone or in combination with other tests) has become both important and topical at the current time.

In Down's syndrome screening programs, the false-positive rate has to be kept low, because a positive result can lead to an invasive procedure with the potential risk for miscarriage. However, in screening for preeclampsia, there is no direct risk to the mother or fetus from the "classification" of being at high risk for the condition. Patients could potentially benefit from increased surveillance. Of course, the anxiety that such a classification may cause should not be underestimated. Nevertheless, it would be reasonable to maximize the detection rate for preeclampsia, at the cost of a higher false-positive rate. Although it is too early at the present time to recommend a national program of screening for preeclampsia in the first trimester, it is important to remember that uterine artery Doppler screening is not difficult to perform, is noninvasive, adds little to the scanning time, is widely available, and can be added to an already scheduled sonographic examination, which is now commonly performed in most hospitals (eg, nuchal translucency).

The sensitivity of uterine artery Doppler velocimetry may also be improved by adding single or multiple biochemical markers to the screening algorithm. In the second trimester, it appears that based on the evidence, a combination of either placental proteins or angiogenic factors or homocysteine with uterine artery Doppler velocimetry slightly improves the predictive ability for preeclampsia compared to the tests alone. However, large, prospective studies are necessary before claims can be made that combinations of tests are effective to predict preeclampsia.

We anticipate and hope that through future studies, further maternal serum markers will be identified, and that the best "panel" of markers along with uterine artery Doppler velocimetry will be established, in order to provide the highest sensitivity and predictive values for preeclampsia, SGA fetuses, etc. Ideally, this would be performed in the first trimester, and such identified high-risk women may be the most likely to benefit from pharmacological intervention in future trials.

Acknowledgments

This research was supported (in part) by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.

The authors would also like to acknowledge and thank Eli Maymon and Gianluigi Pilu for their past contributions to this chapter, which was previously published in the sixth edition.

Highlighted References