ISUOG Consensus Statement on maternal hemodynamic assessment in hypertensive disorders of pregnancy and fetal growth restriction

The International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) is a scientific organization that encourages sound clinical practice, and high-quality teaching and research, related to diagnostic imaging in women's healthcare. The ISUOG Clinical Standards Committee (CSC) has the remit to develop Practice Guidelines and Consensus Statements as educational recommendations that provide healthcare practitioners with a consensus-based approach, from experts, for diagnostic imaging. They are intended to reflect what is considered by ISUOG to be the best practice at the time at which they are issued. Although ISUOG has made every effort to ensure that Guidelines and Consensus Statements are accurate when issued, neither the Society nor any of its employees or members accepts any liability for the consequences of any inaccurate or misleading data, opinions or statements issued by the CSC. The ISUOG CSC documents are not intended to establish a legal standard of care because interpretation of the evidence that underpins them may be influenced by individual circumstances, local protocol and available resources. Approved Guidelines and Consensus Statements can be distributed freely with the permission of ISUOG ([email protected]). Pregnancy presents a unique cardiovascular challenge. Profound maternal hemodynamic adjustments result from a complex interplay between the maternal cardiovascular system and uteroplacental perfusion, which permits fetal development while maintaining healthy maternal homeostasis1-5. Failure to meet this challenge can lead to maternal hypertensive complications (hypertensive disorders of pregnancy (HDP)) and/or fetal growth disorders (fetal growth restriction (FGR))6-9. Monitoring and management of HDP and FGR has been based largely on maternal blood pressure (BP) monitoring and fetal growth assessment10. Promising research offers the prospect of a less reactive and more proactive approach. Maladaptation to the requirements of pregnancy can be detected in a preclinical phase of HDP and FGR by maternal hemodynamic assessment, offering opportunities for targeted hemodynamic interventions11-15. Pregnancy also serves as a stress test for future cardiovascular health, especially if superimposed upon subclinical cardiometabolic or cardiovascular vulnerabilities11, 16-19. This Consensus Statement provides an update on the role of maternal hemodynamic assessment in HDP and FGR. A proposal for this Consensus Statement was submitted to and evaluated by the ISUOG CSC. The authors of this Consensus are members of the International Working Group on Maternal Hemodynamics (https://maternalhemodynamics.com/) and were involved in both the development and the writing stages. Members of the International Working Group on Maternal Hemodynamics who were not involved in the writing of this Consensus were invited to review the document (reviewing group) and those who responded are acknowledged. Stakeholders were not included in the writing process. This process was not intended to represent a systematic review. The preselected topics were divided among the authors, who worked in pairs and performed a literature search. T.S. collated the document and proposed the key points. The final document and key points were revised by all authors. Several meetings were organized to discuss any points on which there was not initial consensus, until complete agreement was reached. The document was then finalized after comments from the reviewing group, the ISUOG CSC and the ISUOG Board of Trustees. The authors of this Consensus Statement are experts in the field of maternal hemodynamics and have expressed their opinion based on the available literature and evidence. Any clinician applying this Consensus is expected to use their independent medical judgement in the context of the individual clinical circumstances to determine patient care. Traditionally, peripheral BP is measured by auscultation using an arm cuff connected to a manometer. Increasingly, automated oscillometric devices, validated for pregnancy, are becoming the recommended standard20, 21. Further developments have allowed additional analysis of the arterial waveform for estimation of central BP and arterial function. Markers of arterial stiffness and compliance, such as pulsed-wave velocity and the augmentation index (AIx)22, 23, help to understand the pathophysiology behind HDP and FGR, predict cardiovascular complications, and evaluate and direct prevention and treatment24-26. Assessment of central hemodynamic factors that determine BP (cardiac output (CO, in L/min) and systemic vascular resistance (SVR, in dyne.s/cm5)) is increasingly accessible using easily operable devices27. Considering the pivotal role of central hemodynamic functioning, this Consensus Statement focuses primarily on CO assessment. CO is the product of stroke volume (SV) and heart rate (HR) (CO = SV × HR). Accuracy of its evaluation is improved by averaging measurements over several heartbeats3-5, to compensate for respiratory variations. After 20 gestational weeks, CO is best evaluated with the woman in a left lateral position, to minimize caval and aortic compression by the gravid uterus28, 29. The heart can rapidly modulate CO in order to meet the demands of pregnancy, with a wide range of normality (4–20 L/min) depending on activity level, age, fitness, body size, gestational age and metabolic rate. While trimester-specific reference ranges exist, device-specific normalization, that takes into account maternal age and posture, does not. Indexing for actual or prepregnancy body surface area or height may be of additional value when comparing measurements between individuals27, 30-32. Individual or separate CO measurements can be of value for instant hemodynamic assessment, while repeated or continuous measurements are appropriate for trend analysis. With knowledge of CO and BP, SVR can be calculated ((SVR = 80 × (MAP − RAP)/CO) in dyne.s/cm5), where MAP represents mean arterial pressure and RAP is right atrial pressure. RAP (or central venous pressure) is usually low in pregnancy (0–5 mmHg) and therefore is usually omitted from the calculation33, 34. Arterial function can also be evaluated by Doppler assessment of peripheral vessels, for example the uterine and/or opthalmic arteries. CO is determined most accurately by invasive measurements using thermodilution or indicator (lithium) dilution, or by cardiac magnetic resonance imaging (MRI). The former require insertion of intracardiac, central venous and/or arterial catheters (e.g. pulmonary artery catheters), limiting their indication to critically ill pregnant women in an intensive care unit setting33, 34 . Cardiac MRI is non-invasive, but its limited availability and high cost restrict its use in an obstetric context28, 35. Transthoracic echocardiography permits non-invasive assessment of CO by measuring the left ventricular outflow tract diameter and pulsed-wave Doppler outflow signal along with HR. It has been validated against invasive measurements in severely ill pregnant women and can be considered an alternative reference technique36, 37. Transthoracic echocardiography enables evaluation of systolic and/or diastolic function (assessing ejection fraction using two-dimensional and/or M-mode imaging, strain and strain rate using speckle tracking, tricuspid annular plane systolic excursion (TAPSE), and mitral inflow using pulse-wave and tissue Doppler imaging) and preload assessment (atrial size and inferior vena cava size and collapsibility)38-40. Another means of assessing CO is with a non-invasive mobile device, using a continuous-wave Doppler transducer placed in the suprasternal notch40. This technique is limited by assumptions regarding the left ventricular outflow tract diameter and acoustic guidance for probe orientation. However, by including more heartbeats in the calculation, the accuracy is enhanced. An inert gas rebreathing technique can be used to calculate CO based on Fick's principle41. A blood-soluble and an insoluble gas are delivered though a closed breathing assembly and relative levels over a few respiratory cycles are measured by a gas analyzer in the mouthpiece. As with pulmonary artery catheters, pulmonary blood flow or right ventricular output is measured, which is usually identical to left ventricular output. In obstetrics, this technique is more suited for research than clinical practice42. With impedance and reactance cardiography, a very-low-amplitude high-frequency current is transmitted between skin electrodes on the thorax. CO can be estimated by calculating differences in impedance (bioimpedance) or phase shifts (bioreactance) induced by changing blood flow during the cardiac cycle43, 44. Pulse-contour analysis estimates CO from a peripheral arterial waveform, which can be obtained invasively or non-invasively33, 45. Both impedance and reactance cardiography and pulse-contour methods are operator-independent and permit continuous monitoring for trend analysis, but their algorithms to calculate CO are based on assumptions that may affect accuracy. Most non-invasive techniques lack validation in pregnancy against reference techniques. The few that have been compared with reference techniques often fail the current rigorous standards for absolute values and trend analysis46. Consequently, CO values obtained with one technique are not comparable with those obtained using another technique, and the techniques cannot be used interchangeably. Still, it would be unwise to disregard the benefits from these devices, as their accessibility permits the gathering of large amounts of valuable hemodynamic information. In addition, using a single device repeatedly in the same patient overcomes the differences between techniques and allows the investigation of changes over time. As such, these techniques may be valuable for monitoring changes as pregnancy progresses. During pregnancy, the mother's body undergoes cardiovascular changes leading to hemodynamic adaptations required to respond to the increasing metabolic and oxygen demands from the uterus, placenta and fetus and to prepare the mother for possible blood loss during delivery (Figure 1). To understand the possible role of maternal hemodynamics in pathological conditions, it is important to understand these changes in healthy pregnancy, as normal maternal hemodynamic adaptation is a prerequisite for development of a normal pregnancy42, 47. Maternal hemodynamic changes can be observed during the menstrual cycle. Compared to the follicular phase, during the luteal phase, the HR, plasma volume (PV), CO and central and peripheral arterial compliance are increased, while the MAP and SVR are decreased48. Similar changes are present early in pregnancy, from 6 weeks of gestation, prior to the development of a placental circulation42, 48, 49. The hypothesized mechanism underlying maternal hemodynamic changes in early pregnancy (Figure 2) is that a fall in SVR, by up to 20–25%50, acts as the trigger event, most likely due to reduced vascular tone, leading to a decrease in MAP and a compensatory increase in CO51. The initial increase in CO is caused by an increase in HR48. These changes are accompanied by renin–angiotensin–aldosterone-system (RAAS) activation due to increases in renal plasma flow and glomerular filtration rate and reduction in MAP48, 51. The result of RAAS activation is sodium and water reabsorption, with consequent PV expansion, hemodilution and increased preload52. The increased preload leads to improved cardiac filling during diastole. Together with decreased cardiac afterload, this results in increased SV and, consequently, in increased CO48. As a result of increased preload, the maternal heart undergoes changes: the left atrial diameter, left ventricular diastolic and systolic dimensions and ventricular wall thickness all increase, resulting overall in increased cardiac mass, surface area (so-called myocardial eccentric hypertrophy) and performance5, 53. Maternal cardiac morphological changes start at around 8 weeks of gestation, but become significant at around 12 weeks of gestation52. The maintenance of CO is also guaranteed by decreased cardiac afterload, which is accomplished by the decrease in SVR, while the normal perfusion pressure is associated with higher arterial compliance54. Concomitantly, increased venous compliance accommodates a higher quantity of PV, contributing to preload increase and further maintenance of CO48. A detailed discussion of molecular pathways is beyond the scope of this document, but it has been hypothesized that hormonal influences, including estrogen and human chorionic gonadotropin secreted by the synciotrophoblast, and other factors (e.g. relaxin, vascular endothelial factor, nitric oxide (NO)) play a role in physiological maternal hemodynamic adaptation early in pregnancy48, which would mean that the earliest maternal hemodynamic changes in pregnancy develop independently from the fetoplacental unit. After the early first-trimester changes, healthy pregnancy is characterized by various hemodynamic adaptations, including further increases in CO and PV, decreases in SVR and MAP, and changes in maternal cardiac function. Increasing cardiac preload and concomitant decreasing afterload result in progressively higher SV and CO in the first trimester of pregnancy. SV increases by 15–25% early in pregnancy and plateaus by about 20 weeks55, 56. In the second half of pregnancy, the increase in SV is reversed, so that near term there is no difference between pregnant and non-pregnant women56. The increased CO is also caused by the higher maternal HR, which increases by 15–30% early in pregnancy and continues to increase gradually until term. The CO increases overall by 20–50% and peaks by mid to late pregnancy, reaching a maximum mean difference of 1.41 L/min between 29 and 35 weeks of gestation55, 56. It remains controversial as to whether CO increases until term57 or decreases late in pregnancy58; inconsistencies in results between studies can be explained by differences in their design and/or methodology, population and other factors. CO returns to prepregnancy values from 2 weeks postpartum1. PV starts to increase early in pregnancy, peaks during the second trimester and changes minimally throughout the third trimester, returning to normal levels after 6 weeks postpartum4. Around one-third of the PV is contained in the arterial compartment, while two-thirds are located in the venous compartment and represent a hemodynamically inactive intravascular volume59. SVR decreases, with an overall 30% decrease early in pregnancy55 and nadir in the late second trimester56. MAP decreases during pregnancy until 24 weeks of gestation, because the increase in CO is not sufficient to balance the fall in SVR. Both systolic and diastolic BP decrease; however, the reduction in diastolic BP is greater than that in central and brachial systolic BP49. There is also a significant fall in AIx (i.e. arterial stiffness), which explains the reduced central systolic BP. Central systolic BP reflects better than brachial systolic BP the uterine artery pressure, which contributes to the uteroplacental perfusion49. The BP returns to baseline levels during the third trimester55, 56. Increased preload and concomitant decreased afterload have a positive effect on maternal cardiac function in the first trimester of pregnancy. The change in left ventricular diastolic function in the first trimester appears to be determined by an increased contribution of passive left ventricular filling4. In advanced gestation, left ventricular diastolic function seems to worsen progressively, as reflected by a decreased E-wave to A-wave (E/A) ratio4. Morphological changes start in the first trimester and continue throughout pregnancy, peaking around 30 weeks. The left ventricular mass increases by approximately 50%55. Myocardial eccentric hypertrophy observed during healthy pregnancy represents a reversible and uncomplicated phenomenon that has usually resolved by 6 months postpartum60, 61. More profound maternal hemodynamic changes have been reported in uncomplicated multiple compared with singleton pregnancies, including higher CO and lower SVR across gestation62-64. The greater increase of circulating blood volume and the higher concentrations of vasodilating hormones, such as prostaglandins and progesterone, are thought to drive these changes65. In uncomplicated twin gestation, changes in maternal systolic and diastolic function have been documented from the first to the third trimester66-68. In particular, diastolic and mean BP show a U-shaped trend through gestation, reaching their nadir at midgestation. Based on serial echocardiograms, left ventricular systolic function, expressed either by classical parameters (ejection fraction or fractional shortening) or by tissue Doppler parameters (longitudinal contractility, quantified by S′ (myocardial velocity during isovolumetric longitudinal contraction)) worsens progressively during pregnancy62, 67. Moreover, relative diastolic dysfunction is observed, as evidenced by a progressive reduction in the mitral E-wave and an increase in the A-wave velocities, together with consistent modifications of the tissue Doppler parameters (reduction in E′ and increase in A′)67. These findings indicate that the maternal heart is subjected to profound adjustments during the course of a twin gestation, regardless of whether there are cardiovascular complications. The progressive decrease of ventricular performance may depend on the volume overload65, and, in the late stages of apparently uneventful twin gestation, may lead to increased end-systolic residual volumes in the ventricle, thus contrasting with the early diastolic phase and promoting active atrial contribution to left ventricular filling across the atrioventricular valves. In other words, the late stages of an uncomplicated twin pregnancy at the level of the maternal heart may mimic diastolic dysfunction as observed in the early stages of chronic cardiac insufficiency. It is noteworthy that similar findings, suggestive of early diastolic dysfunction, have been described at postpartum echocardiography in women who developed preterm or severe pre-eclampsia (PE) in a singleton gestation17, 69. At second-trimester maternal echocardiography, the left ventricular mass is increased in twin pregnancies compared with singletons, but there are only subtle differences in systolic and diastolic left ventricular functional indices70. The pattern of cardiac changes is similar between monochorionic and dichorionic twin pregnancies. These cardiovascular changes resemble those seen in singletons in late gestation and suggest physiological remodeling in response to the increase in volume loading rather than decompensation of the maternal cardiovascular system. Changes in maternal cardiac function in uncomplicated twin gestation, from early to late pregnancy, include an increase in CO and a parallel decrease in SVR. It is unclear how these cardiac findings differ with respect to chorionicity. Using serial transthoracic echocardiography, one study reported lower maternal CO in second-trimester monochorionic vs dichorionic twin pregnancies, possibly because of the smaller increase in the circulating blood volume66. In contrast, another study recently showed a higher CO in the second trimester in monochorionic compared to dichorionic twin pregnancies68. Most cardiac parameters revert towards the values of early pregnancy soon after delivery, with the exception of some indices of systolic function, such as fractional shortening, ejection fraction and tissue Doppler S′, which have been reported to remain significantly below those observed in the first trimester67. This may indicate that the workload caused to the maternal heart by the increased circulating volumes may still affect the ventricular contractility in the early postpartum period and that complete recovery is delayed. The first classification of PE phenotypes was based on the timing of clinical disease onset, before ('early-onset') or after ('late-onset') 34 weeks of gestation71, 72. Early-onset PE typically follows a more aggressive course, presenting with severe symptoms of maternal organ dysfunction or with poor fetal growth that precedes the development of maternal hypertension, in association with serum markers of immune and endothelium activation73. The maternal hemodynamic characteristics are markedly different between the two phenotypes, both in the clinical14 and in the latent14, 74 phase of the disease15. A new classification system has been proposed that is based on maternal hemodynamics and fetal growth14, 15, 75, 76. PE associated with FGR is characterized by low CO and high SVR, in both preclinical and overt phases of the disease. PE associated with normal fetal growth is not characterized by low CO, while the SVR in the preclinical phase of the disease is low and may increase in the overt phase. This classification allows discrimination between two PE phenotypes, one characterized by vascular-resistance-induced hypertension ('resistance-dominant HDP') and the other by intravascular-volume-induced hypertension ('volume-dominant HDP') (Figure 3). As such, the fundamental physiological definition of BP as the product of CO and SVR (Ohm's law) is pertinent: hypertension is the result of either increased SVR, increased CO or both6, 77. The theory that the placenta is the main etiological driver of PE is challenged by the observation of subclinical abnormal hemodynamic function already present in the preconception stage in women who subsequently developed PE47, 78. An increasing number of clinical, epidemiological and experimental observations support that the abnormal placentation process associated with PE is a result of, rather than a cause of, abnormal maternal hemodynamic function7. Latent suboptimal hemodynamic function before conception predisposes not only to gestational complications, but also to early-onset cardiovascular disease in later life79-81. Moreover, abnormal intrauterine hemodynamics may be a direct link between poor fetal growth and early-onset adult cardiovascular and renal disease82. While impaired maternal cardiovascular performance plays a role in the poor development of placentation, we should not overlook the contribution of immune-system dysfunction, which is observed when defective trophoblast invasion occurs in early gestation83. The association between confirmed poor maternal hemodynamics and the immune system should be a focus of future research84. The classification of PE into resistance-dominant or volume-dominant phenotypes challenges the reported longitudinal hemodynamic evolution from early pregnancy until term in women who eventually develop PE6. In all studies reported since 1990, a hemodynamic imbalance between CO and SVR was present at all stages of pregnancy in women who developed PE8, 78, 85. Early-onset PE typically presents with first-trimester increased SVR and a blunted rise of CO85. Late-onset PE, on the other hand, starts with high CO and normal SVR8. This imbalance can remain unchanged8, 57, or convert from a high-CO circulation to a high-resistance circulation with advancing gestation8, probably due to endothelial damage and intravascular overload86, thereby establishing a 'vicious circle' that, in association with hypoalbuminemia and impaired cerebral blood flow autoregulation, strongly predisposes to eclampsia87. Gestational hypertension (GH) also presents with both volume-dominant and resistance-dominant phenotypes8, 85. Contrary to HDP associated with renal damage, abnormal maternal venous Doppler characteristics are absent in GH6, raising the question as to whether PE-related symptoms of organ dysfunction may relate to venous congestion induced by increased intravenous pressure, reduced venous capacitance or external venous compression due to raised intra-abdominal pressure88. Overall, this evidence supports the abandonment of two traditional concepts of obstetrics. Firstly, the mechanistic time-dependent classification into early-onset and late-onset PE has no clinical relevance, and is meaningful only from an epidemiological perspective. Secondly, the very term 'pre-eclampsia' no longer bears any clinical meaning once we recognize the truly different pathophysiological pathways, which simply share similar downstream symptoms. Thus, PE and GH should be considered under the same umbrella, that is, HDP, with different clinical facets. Accordingly, herein, we use the broader term, 'HDP', unless otherwise indicated (e.g. when discussing studies focused only on PE). Throughout healthy pregnancy, maternal CO correlates positively with fetal growth velocity and birth weight, and the physiological reduction in SVR is associated with increased birth weight. This is in line with the observation of an interaction between maternal cardiac and placental function89; the decrease in SVR coincides with the reduction of resistance in the uteroplacental circulation. Moreover, there is a significant correlation between fetal Doppler indices and maternal hemodynamics. Maternal SVR, CO, heart contractility and dynamicity of circulation all seem to influence the blood flow rate in the umbilical vein90. The evidence of a correlation between maternal, placental and fetal circulations91 highlights the importance of hemodynamic assessment in cases of inadequate fetal growth. The most common cause of FGR is so-called uteroplacental insufficiency, resulting from defective placentation, and, traditionally, FGR has been defined as a 'placental syndrome'92, 93. The poorly functioning placenta restricts nutrient supply to the fetus and prevents normal fetal growth. This condition is characterized by defective trophoblast invasion, abnormal uterine artery adaptation and increased placental vascular resistance. A new interpretation of the pathophysiology of FGR has stemmed from the premise that placental function is dependent on its perfusion15, 93. Thus, adequate systemic cardiovascular adaptation in pregnancy is crucial to ensure optimal placental function. Large population studies have identified the occurrence of FGR as an important risk factor for the development of cardiovascular disease later in a woman's life, with the highest risk in cases of the most severe and early FGR94, 95, regardless of the simultaneous presence of PE. The definition of FGR is based on fetal biometry and Doppler evaluation, and it is classified as early or late according to whether it presents before or after 32 weeks of pregnancy, respectively96. Early FGR has low incidence (0.5%) and is characterized by high-impedance uterine circulation, frequent association with HDP, high risk of perinatal mortality and morbidity, and poor neonatal outcome due to prematurity97. While it is known that defective trophoblast invasion with subsequent failure of spiral artery adaptation is associated with early FGR92, maternal hemodynamic assessment provides further insight, linking evidence of uteroplacental dysfunction with signs of maternal cardiovascular impairment in pregnancies with early FGR. Women with a prepregnancy or early-pregnancy abnormal maternal hemodynamic profile, with low CO, high SVR and a hypovolemic state, have an increased risk of developing FGR47, 98, and this hemodynamic profile is maintained during the pregnancy. These hemodynamic alterations can be identified in apparently healthy women at different preclinical stages before manifestation of the disease, and SVR seems to be the most important predictive factor99, 100. The fact that in patients with chronic hypertension there is an increased risk of FGR further supports the hypothesis that the maternal cardiovascular system plays a crucial role in the pathogenesis of abnormal placental function99, 101. Compared with early FGR, late FGR is less closely associated with HDP, abnormal placental development and spiral artery remodeling. The main clinical challenge is its correct identification and diagnosis, firstly, due to the subtle clinical manifestations and, secondly, due to the fact that FGR might also exist in fetuses with size (abdominal circumference (AC) or estimated fetal weight (EFW)) > 10th percentile97, 102, 103. However, late FGR, with a prevalence of around 5–10%, is much more common than early FGR and thus has an important clinical impact97. The association between late FGR and placental histopathological lesions is less evident than that in early FGR, and the differences compared with healthy pregnancy are often very subtle104. Late FGR, compared with early FGR, shows less pronounced changes in uterine artery impedance92. In about one-third of all pregnancies, there is a worsening of uterine artery Doppler in the late third trimester105, suggesting that this late variation in uterine vascular resistance is not related to trophoblast invasion but is an effect of changes that involve the maternal cardiovascular system106. In these cases, there is a higher risk of low birth-weight percentile105. This observation explains the low predictive capacity of uterine artery Doppler velocimetry to identify late FGR and suggests the importance of maternal hemodynamic assessment in these cases. The second trimester may be optimal for maternal cardiovascular assessment for identification of patients at risk of developing late FGR13. In pregnancies complicated by late FGR, maternal hemodynamic parameters do not show physiological cardiovascular adaptation, but show a flat, static pattern, with low CO and high SVR106. When fetal size (AC or EFW) is below the 10th percentile, a priority is to differentiate between small-for-gestational age (SGA) and FGR. Some SGA fetuses might be constitutionally small but otherwise healthy and, consequently, at lower risk of adverse out