Statement on Pregnancy in Pulmonary Hypertension from the Pulmonary Vascular Research Institute

Mechanical and hormonal changes in pregnancy

Mechanical changes in pregnancy. Mechanical changes in pregnancy are primarily due to hormone-mediated changes in maternal chest wall diameter and enlargement of the uterus, thereby affecting cardiac and pulmonary function. In addition, in the supine position, blood flow in the inferior vena cava may be compromised because of uterine compression.

Hormonal changes in pregnancy

Sex hormone levels increase dramatically during pregnancy and have a major effect on the body.^26–30 The fetoplacental unit formed by the fetus and the placenta produces all the hormones necessary for a successful pregnancy. Steroid precursors are delivered from both the fetus and the mother.

One of the first hormones to peak during pregnancy is plasma human choriogonadotropin (hCG). This peak occurs in the tenth week of pregnancy. One of the many effects of hCG is the stimulation of the hormone relaxin, which, like estrogens and progesterone, mediates vasodilatory responses and facilitates blood flow to critical organs. ³¹

Progesterone and estrogens are the major sex hormones for the remainder of the gestation period. Plasma progesterone is initially produced by the corpus luteum, but its production is taken over by the placenta during the transition to the second trimester. Progesterone exhibits a first peak in the first month of pregnancy, followed by a sharp and continuous increase throughout the remainder of the pregnancy, ending with a peak concentration at around 340 pg/mL immediately before delivery.^27,30 The major role of progesterone is to create a protective environment for the fetus in the uterine cavity by stimulating endometrial glands to nourish the zygote and by maintaining the homeostasis of decidual cells. Furthermore, progesterone inhibits the occurrence of uterine contractions by decreasing prostaglandin synthesis and by attenuating sensitivity to oxytocin. Progesterone also mediates major cardiovascular and respiratory adaptations in pregnancy (see below).

Levels of the three major estrogens (estrone [E1], 17β-estradiol [E2], and estrione [E3]) rise steadily throughout pregnancy. E2 levels reach values of up to 7,200 pg/mL.^27,30 Major effects of estrogens include stimulation of the growth of the myometrium and of the ductal system of the breast. Like progesterone, estrogens also have significant effects on the cardiopulmonary system (reviewed below).

Other sex hormones increased during pregnancy include prolactin, sex hormone–binding globulin, dehydroepiandrosterone (DHEA), and testosterone.^26–29 Both progesterone and estrogen levels drop sharply immediately after delivery.

Pregnancy effects on the cardiovascular system

Plasma volume and dilutional anemia. One of the most significant changes in the cardiovascular system is a significant increase in plasma volume. Plasma volume starts to increase early in pregnancy and peaks shortly before delivery at values that are 50%–70% above the patient's prepregnancy values.^32–34 The major cause of the pregnancy-induced increase in plasma volume is hormone-mediated vasodilation, which leads to underfilling of the vascular system and atria, resulting in activation of the renin-angiotensin-aldosterone system and a decrease in natriuretic peptide release, allowing for significant retention of sodium and volume (6–8 L).^35–38 The latter is mainly contained in the uterus, breast, muscle, and skin. Red cell mass increases above the prepregnancy baseline, also as a consequence of increased erythropoietin production, but only by 20%–30% (and even less so in patients not taking iron supplements).^32,39,40 Since red cell mass rises less than the plasma volume, a significant dilutional anemia occurs, with normal hemoglobin values as low as 10.5 g/dL. ⁴¹ The resultant decrease in blood viscosity facilitates perfusion of the fetoplacental unit and may also lower cardiac work.

Cardiac axis, heart size, and heart rhythm. Because of mechanical forces from the enlarging uterus and a resulting upward shift of the diaphragm, a leftward shift of the cardiac axis occurs. Consequentially, nonspecific ST and T wave changes can be seen on an electrocardiogram. ⁴² This is accompanied by mild dilation and stretching of all cardiac chambers, as well as mild tricuspid regurgitation,^43,44 changes that are due to the increase in plasma volume. Left ventricular (LV) mass increases slightly;^44,45 however, the ratio of wall thickness to ventricular radius does not change, indicating an eccentric hypertrophy similar to what is seen with athletic training. ⁴⁶ These shifts in cardiac axis and anatomy, in combination with hemodynamic, hormonal, and autonomic changes, represent the major causes of the frequently observed benign supraventricular arrhythmias during pregnancy. ⁴⁷ Similarly, increases in heart rate by 15–20 beats per minute are common.^48,49

CO, systemic vascular resistance, and mean arterial pressure (Fig. 1). CO rises sharply in the first trimester and then increases more gradually to peak values of around 30%–50% above the prepregnancy baseline.^49–52 Initially, this is mediated by an up to 35% increase in stroke volume (SV). The reason for the increase in SV is likely multifactorial and due to the increased plasma volume, decreased afterload, increased contractility, and eccentric hypertrophy associated with increased LV compliance.^46,53 The increase in contractility likely results from a Frank-Starling response triggered by increased plasma volume and possibly also from direct inotropic effects of sex hormones (e.g., estrogens and prolactin). However, increases in contractility have not been uniformly described, and transient decreases in systolic function have been reported as well.^54,55 Later in pregnancy, an additional increase in heart rate contributes to the increases in CO as well. CO values of up to 8.7 L/min at week 36–38 may be seen.^45,56–58 LV ejection fraction remains unchanged.^43,44 B-type natriuretic levels increase slightly but remain within normal limits. ⁵⁹

OPEN IN VIEWER

Figure 1.

Changes in cardiac output (CO), mean arterial pressure (MAP), and systemic vascular resistance (SVR) during pregnancy. Values above dotted line represent an increase from baseline, while values below dotted line depict a decrease.

Systemic vascular resistance (SVR), on the other hand, decreases by up to 40% as a consequence of progesterone- and estrogen-mediated vasodilation as well as the generation of a high-flow and low-resistance circuit in the uteroplacental circulation; this results in a significant drop in diastolic blood pressure.^49,60–64 The mechanisms of hormone-mediated vasodilation are not completely understood but likely include increased endothelial prostacyclin and nitric oxide production and decreased responsiveness to the vasoconstrictive effects of angiotensin II and norepinephrine, as well as reduced aortic stiffness.^65–70 The latter allows for increased compliance of the vascular system. The structural correlates of these changes include fragmentation of reticular fibers and changes in the structure of elastic fibers, as well as hypertrophy and hyperplasia of smooth muscle cells. ⁷¹

Because of the opposing effects of increased CO and decreased SVR on mean arterial pressure (MAP), the latter decreases only slightly. Systolic and diastolic blood pressure may decrease slightly in the first 2 trimesters (to a statistical mean of about 105/60 mmHg); however, in the third trimester these values usually increase back to prepregnancy baselines, as does MAP.^49,60–64

Pulmonary vasculature (Fig. 2). Changes similar to those in the systemic vasculature are observed in the pulmonary vascular system, where the increase in CO and the decrease in PVR result in a grossly unchanged mPAP. As in the systemic vasculature, the decrease in PVR is primarily hormone mediated; recruitment of previously nonperfused pulmonary vessels (similar to what is seen during exercise) has been suggested as well. A normal pulmonary vascular system can therefore easily accommodate the increased CO and increased pulmonary blood flow that occur during pregnancy (Fig. 2; upper-right corner). However, in the setting of pulmonary vascular disease (e.g., PAH or pulmonary embolism), these compensatory mechanisms are significantly impaired, resulting in a significant afterload stress on the RV (Fig. 2; lower-right corner). The inability of the pulmonary vasculature to accommodate the increased CO therefore explains the frequent worsening of PH and the frequent occurrence of RV failure during pregnancy or in the postpartum period, where CO rises even more (see below).^21–23

OPEN IN VIEWER

Figure 2.

Adaptation of the pulmonary vascular system and the right ventricle (RV) to increased pulmonary blood flow during pregnancy in a healthy patient and in pulmonary vascular disease. Note that the diseased pulmonary vasculature in pulmonary arterial hypertension (PAH; characterized by vasoconstriction, pulmonary vascular remodeling with lumen obliteration, and in situ thrombosis) is unable to accommodate the increased cardiac output, thus leading to RV strain, dilation, and eventually decompensation. PE: pulmonary embolism.

Effects of labor on the cardiovascular system. While the first 9 months of pregnancy are characterized by marked but gradual changes in the cardiovascular system, labor and delivery are associated with even more rapid and dramatic changes. In particular, the intrathoracic pressure swings and volume shifts associated with labor and delivery have significant effects on cardiac preload and afterload. For example, the pronounced respiratory efforts associated with labor and delivery cause significant changes in intrathoracic pressure, thus leading to marked pressure swings in the central venous system as well as in the arterial system. ⁷² Furthermore, large volume shifts occur during and after delivery through the release of vena caval obstruction, as well as through autotransfusion of 300–500 mL of blood from uterine contractions.^39,73 A significant additional increase in blood volume is therefore seen in the peripartum and postpartum periods, accompanied by increases in CO (15%–25% above prelabor values) and SVR.^39,73 On the other hand, postpartum hemorrhage may lead to marked volume shifts in the other direction, highlighting the complexities of labor and delivery in patients with pulmonary vascular disease.

Pregnancy effects on the respiratory system

Lung mechanics and gas exchange. Early in pregnancy, structural changes occur in the thorax to accommodate the enlarging uterus. The circumference of the lower chest wall increases, and the subcostal angle may increase from 68° to 100° or more in the first trimester. The chest diameter increases by a minimum of 2 cm, resulting in a “barrel chest” appearance of the thorax.^74–76 These changes are due to the relaxing effects of the hormone relaxin on the ligamentous attachments of the lower ribs. ⁷⁷ By term, the diaphragm rises up to 4 cm. Interestingly, diaphragmatic excursion is not limited by the enlarging uterus and increases by up to 2 cm.

Pregnancy is a hypermetabolic state, and oxygen consumption increases by up to 20%. ⁷⁸ Respiratory adaptations to the increased oxygen demand are primarily mediated via progesterone-mediated increases in tidal volume, while respiratory rate does not change significantly.^79,80 The underlying mechanisms of progesterone-mediated hyperventilation are not completely understood; increased sensitivity of the medulla to carbon dioxide has been suggested, as well as direct stimulatory effects of progesterone on the respiratory center.^79,81,82 Minute ventilation increases by approximately 50%, resulting in a respiratory alkalosis that is accompanied by a compensatory metabolic acidosis.^79,83–85 The consequence of the increased alveolar ventilation is a significant rise in alveolar and arterial oxygen tension^84,86 and also a sense of dyspnea. In fact, up to 70% of normal pregnant women experience dyspnea during pregnancy. ⁸⁷ The changes in maternal arterial oxygen and carbon dioxide tensions (Pao₂ and Paco₂, respectively) facilitate oxygen and carbon dioxide transfer across the placenta. Normal values for pH, Pao₂, and Paco₂ in pregnancy are provided in Table 2.

OPEN IN VIEWER

Table 2.

Normal values for pH and arterial oxygen and carbon dioxide tensions (Pao₂ and Paco₂, respectively) in pregnancy

Measure	Normal value in pregnancy
pH	7.39–7.45
Paco2, mmHg	25–33
Pao2, mmHg	92–107

Sources: Handwerker et al.; ⁸³ Fadel et al.; ⁸⁴ Lim et al.; ⁸⁵ Spiropoulos et al. ⁸⁶

Functional residual capacity (FRC) decreases by 20% as a consequence of displacement by the enlarging uterus. ⁸⁸ This is the correlate for the rapid desaturations that occur in pregnant patients despite the slightly increased Pao₂.

Airway anatomy. Upper-airway anatomy is affected by pregnancy-induced weight gain and by pregnancy-induced fluid retention and associated oropharyngeal edema; this has significant implications for airway management in pregnant patients, as a difficult airway situation may arise. This is exemplified by a study demonstrating an increase by 34% in the number of patients with a Mallampati class 4 airway from week 12 to week 38. ⁸⁹ In addition, the process of labor may significantly affect airway anatomy. Specifically, a recent study demonstrated that the modified Mallampati classification increased by 1 grade in 33% of parturients and by 2 grades in 5% of parturients. ⁹⁰ Interestingly, airway edema does not correlate with the duration of labor or the amount of fluid administered and may still be present, at least in part, at 48 hours postpartum. ⁹⁰ The exact cause of this phenomenon is unclear. The combination of a difficult airway and decreased FRC can make endotracheal intubation extremely challenging, a scenario that is especially precarious if the patient has an already compromised pulmonary vascular bed, as is the case in PH.

Physiologic changes in the postpartum period

The postpartum period is characterized by a rapid drop in estrogen and progesterone levels, followed by gradual normalization over 6 weeks. Blood volume and hematological changes normalize within 6–8 weeks, ⁹¹ while CO and SVR may take up to 3 months, or even longer, to reach prepregnancy values. ⁹²

Summary

Pregnancy effects on the pulmonary vasculature in PAH

As reviewed in “Overview of PH classification and pathophysiology,” the normal pulmonary vasculature responds to pregnancy by dilating and by recruiting previously nonperfused vessels, resulting in decreased PVR. Along those lines, the impact of vasoconstrictor stimuli is attenuated, as evidenced by the observation that pregnancy is characterized by attenuated hypoxic vasoconstriction.^93,94 Women have a more distensible pulmonary circulation than men; ⁹⁵ however, whether these changes are even more pronounced during pregnancy is unknown.

While it is unclear how many of the changes observed in the pulmonary vasculature during pregnancy are directly attributable to increases in sex hormone levels, there is irrefutable evidence that sex hormones have many pharmacological influences on the pulmonary circulation. For example, E2, testosterone, and progesterone all attenuate PA vasoconstriction.^96–100 Underlying mechanisms include increases in endothelial nitric oxide synthase activity, increased prostacyclin synthase activity, and decreased endothelin 1 activation.^96,101–106 Interestingly, progesterone receptors have been identified in myofibroblasts resident in plexiform lesions. ¹⁰⁷ Progesterone causes pulmonary vasodilation ⁹⁹ and may inhibit vascular remodeling and attenuate experimental PH. ¹⁰⁸

The effects of endogenous E2 in PH are largely unknown but likely to be complex. For example, women are more prone to developing PAH,^13,24 but E2 levels and female sex correlate with higher RV ejection fraction,^109–111 and female sex is associated with better PAH survival.^13,24 E2 has several interesting properties relevant to PAH pathogenesis. Particularly, in addition to vasodilator effects, it has proangiogenic and anti-inflammatory properties. ¹¹² Interestingly, pulmonary microvascular density, markers of angiogenesis, and diffusion capacity in women vary over the menstrual cycle, and estrogen replacement in ovariectomized animals increases lung capillarization, thus suggesting that angiogenesis and diffusion characteristics in pregnant patients may be influenced by E2 and other sex hormones as well. ¹¹³ Depending on the context, E2 or its metabolites may exert antiproliferative or proproliferative effects on vascular endothelial and smooth muscle cells.^114–120 The antiproliferative mechanisms have been implicated in protective E2 effects in animal models of PAH;^{105,106,112,115,116} whether they play a role in the adaptation of the pulmonary vasculature to pregnancy, however, remains unknown.

DHEA exerts a variety of effects on the pulmonary vasculature that include vasodilation and stimulation of antiproliferative pathways;^121–126 increases in this hormone thus would be expected to facilitate the adaptation of the pulmonary vascular system to pregnancy-induced hypervolemia. Major effects of sex hormones on the pulmonary vasculature are depicted in Figure 3.

OPEN IN VIEWER

Figure 3.

Pathophysiology of pulmonary hypertension (PH) and right ventricular (RV) dysfunction in pregnancy. Items in red represent the underlying preexisting alterations in PH; items in blue represent pregnancy-related modifiers that may aggravate these alterations. Potential contributions of sex hormones are shown in yellow (derived from studies in normal animals as well as PH models). Decreased pulmonary vasodilation and vascular recruitment, as well as hypervolemia and volume shifts, are the major contributors to RV dysfunction and failure in pregnant PH patients. Volume shifts are primarily due to intermittent mechanical compression of the inferior vena cava, effects of respiration-induced changes in intrathoracic pressure on venous return, autotransfusion during and after delivery, and postpartum blood loss. Hypotension, hypoxemia, and arrhythmias may further worsen RV failure and initiate a vicious cycle. DHEA: dehydroepiandrosterone. ^*Whether worsening pulmonary vascular remodeling contributes to worsening of PH and RV function during pregnancy remains unknown. ^#The contribution of 17β-estradiol (E2) to pulmonary vascular remodeling in PAH is unknown, and inhibitory as well as stimulatory effects may exist, depending on the clinical context.

Relaxin has vasodilator properties in the systemic vasculature^31,127 and appears to act as a pulmonary vasodilator as well.^127–129 It is unknown whether the structural changes that are reported in the systemic vasculature of pregnant subjects (e.g., remodeling of the extracellular matrix, with changes in the structure of reticular and elastic fibers ⁷¹ ) are also present in the pulmonary vascular compartment, but a report of relaxin-induced inhibition of collagen accumulation in the main PAs in a PH model suggests that processes similar to those in the systemic vasculature may also occur in the pulmonary vascular component. ¹²⁹

Several other hormones implicated in the pathogenesis of PAH are further increased by pregnancy. For example, increased prolactin may play a role in precapillary PAH. ¹³⁰ Specifically, this hormone is already significantly increased in PAH patients and correlates inversely with the 6-minute-walk distance and peak oxygen uptake. Serum parathyroid hormone can also be elevated in PH. ¹³¹ Elevated plasma aldosterone levels were reported in PAH patients. ¹³² Finally, pregnancy increases glucocorticoid levels and may lead to hyperglycemia. In addition, maternal diabetes is a risk factor for PPHN. ¹³³ Hence, pregnancy can upregulate many hormones already pathogenically elevated in PAH. It is possible that increases in these pharmacologically active hormones may exacerbate their deleterious effects on the pulmonary vasculature.

The overarching problem in the pregnant PH patient is that the physiologic compensatory vasodilator response of the pulmonary vasculature is decreased or absent, thus leading to a significant increase in mPAP and PVR.^8–10 The inability of the pulmonary vascular bed to accommodate the increased CO results in significant RV afterload stress, and RV strain and RV failure may occur (Fig. 3). The resulting hypotension and hypoxemia (the latter potentially exacerbated by increased right-to-left shunting through a patent foramen ovale) represent major health risks for both the mother and the fetus; this is reflected by the dramatic morbidity and mortality associated with pregnancy in PAH.^21–23 Deterioration is most frequent between weeks 20 and 24, during labor and delivery, or in the postpartum period,^21–23 as these time points represent the stages with the most significant changes in hemodynamics and volume status. Deterioration around week 24 reflects an inability of the cardiopulmonary system to accommodate the increased CO and meet cardiovascular demands, while deterioration during labor and delivery or in the postpartum phase frequently is triggered by volume shifts and intravascular pressure swings, as well as the negative effects of pain, Valsalva maneuvers, hypoxia, and acidosis on pulmonary vascular tone and/or RV function (Fig. 3). Thromboembolic phenomena may contribute to deterioration in the peripartum period as well. A recent review of the literature demonstrated that the majority of maternal deaths occurred in the peripartum period, mainly within the first month of delivery, with RV failure and circulatory collapse being the main causes of death. ²¹

Since the pulmonary vasculature in PAH is characterized by nitric oxide and prostacyclin deficiency as well as by increased endothelin 1 activation,^8,10 there appears to be a lack of substrate for sex hormones to initiate their vasodilator properties. Furthermore, there is concern that prosurvival and proproliferative effects of estrogens may worsen the pulmonary vascular remodeling in pregnant PAH patients. For example, several studies have demonstrated proproliferative effects of E2 on PA smooth muscle cells (PASMCs) in animal models of PH and in human disease.^134–137

In premenopausal women, E2 synthesis occurs mainly in the ovarian follicles and corpus luteum; synthesis also occurs to a lesser extent in nonglandular tissues such as adipose tissue, liver, skin, muscle, and brain. In men and postmenopausal women, adipose tissue is a major source of estrogen synthesis. ¹³⁸ Aromatase (CYP19A1) is a member of the cytochrome P450 (CYP) superfamily and synthesizes E2 through the aromatization of androgens, specifically testosterone and androstenedione. Further metabolism is mediated by several CYP enzymes. In particular, CYP1B1 catalyzes the oxidation of E1/E2 to 2- and 4-OHE2 (2- and 4-hydroxy estradiol, respectively). E2 hydroxylation can occur by CYP enzymes (including CYP1B1), resulting in 6, 7, 15, and 16 (α- and β-) hydroxyestradiol. E2 can also be converted to (1) E1 by 17β-hydroxysteroid dehydrogenase and hence via oxidation to 16α-OHE1, which is mitogenic, or (2) 2-OHE2 by CYP1A1/2, CYP1B1, and CYP3A4. Catechol-O-methyltransferase (COMT) catalyzes the methylation of catechol-E2s to methoxy-E2s, which simultaneously lowers the potential for DNA damage and increases the concentration of 2 methoxyestradiol, an antiproliferative metabolite. Both aromatase and CYP1B1 are expressed by human PASMCs and by the medial layer of human PAs; notably, expression of aromatase is increased in female PASMCs.^139,140 Hence, there is local synthesis and metabolism of E2 in human PAs. It has also recently been shown that endogenous E2 contributes to the development of PH in the female hypoxic mouse and the sugen/hypoxic rat, in part by restoring impaired bone morphogenetic protein type 2 receptor (BMPR2) signaling in the female lung. ¹³⁹ This is significant, as the primary genetic defect of heritable PAH (HPAH), present in at least 70% of cases of HPAH, is a mutation in the gene encoding BMPR2, coupled with dysfunctional Smad signaling that leads to abnormal cell proliferation associated with pulmonary vascular disease. ¹⁴¹ Studies in human PASMCs have also demonstrated that E2 enhances other mediators of proproliferative signaling, such as the mitogenic transcription factors CEBPβ and FOS, ¹³⁷ the calcium-binding protein S100A4/Mts1, ¹³⁵ and several components of the serotonin production and transport axis (all of which have been implicated in PAH pathogenesis). ¹³⁶ A contribution of proproliferative estrogen metabolites (e.g., 16α-OHE1) to PAH development has recently been recognized,^140,142 and it is possible that excess production of such metabolites during pregnancy may contribute to worsening pulmonary vascular remodeling and hemodynamics in PAH patients. Pregnancy exerts complex effects on the various CYP isoforms; ¹⁴³ however, whether and how pregnancy affects CYP1B1 expression and activity remain unknown. In addition to genetic alterations in CYP1B1, two promoter single-nucleotide polymorphisms in the gene coding for aromatase that result in elevated E2 production have been associated with an increased risk of portopulmonary hypertension. ¹⁴⁴

Knowledge gaps. While it is plausible that estrogens and other sex hormones promote pulmonary vascular remodeling in PAH, it is unknown whether the frequently observed worsening of PAH in pregnancy is indeed due to direct effects of sex hormones on the pulmonary vasculature or merely reflects an effect of the hypervolemia and volume shifts that are associated with the pregnant state. Studies investigating whether pulmonary vascular remodeling indeed progresses in pregnant PAH animals are therefore needed. Along those lines, the effects of sex hormones on pulmonary vascular remodeling during pregnancy require further investigation. Similarly, the contribution of estrogen metabolites to PAH development in pregnancy requires further study. Since worsening of PAH frequently occurs in the postpartum period^21–23 and therefore at a time point at which sex hormone levels decrease dramatically, it is possible that a “sex hormone withdrawal phenomenon” results in PA vasoconstriction in the postpartum state. However, this has not yet been investigated. Studies investigating cyclical changes in vasomotor tone related to the menstrual cycle demonstrate worsening vasoconstriction during periods of low estrogen levels,^{100,145–147} suggesting that such a phenomenon may contribute at least in part to the frequent PAH decompensation observed in the postpartum stage.

Pregnancy effects on the RV in PAH

During pregnancy, both the LV and the RV exhibit hypertrophic growth in order to compensate for the increase in blood volume and CO. ⁵⁴ In particular, these chambers undergo eccentric hypertrophy, a finding that is characterized by a proportional change in wall thickness. ¹⁴⁸ Even though diastolic function may be impaired,^44,54 the hypertrophy is considered to be physiological (or adaptive), similar to the cardiac hypertrophy observed in athletes.^148,149 While the SV increases, LV systolic function is maintained,^43,44 although transient decreases in systolic function have been described.^54,55 On the other hand, other studies found no change in LV ejection fraction.^43,44 The histological correlate of the increase in heart size and weight is cardiomyocyte hypertrophy, with an increase in cardiomyocyte length. ¹⁴⁸ Pregnancy-induced cardiac hypertrophy is reversible and starts to regress soon after delivery.^50,54

Cardiac function and remodeling during pregnancy. While the effects of sex hormones and pregnancy on the LV have been relatively well characterized, these data cannot necessarily be extrapolated to the RV, as the two chambers are embryologically, anatomically, physiologically, biochemically, and electrophysiologically different.^150,151 However, studies specifically investigating the impact of sex hormones and pregnancy on the RV are sparse, and investigations are needed that specifically focus on the effect of these factors on RV structure and function. Studies from the LV have demonstrated that both mechanical stress associated with pregnancy-induced hypervolemia and direct effects of sex hormones cause a wide range of changes in the myocardium. ¹⁵² These include the activation of second messengers such as protein kinase C, mitogen-activated protein kinase, and Src tyrosine kinase. ¹⁴⁸ It has been proposed that activation of the latter represents one of the mechanisms of estrogen-mediated heart hypertrophy in late pregnancy. ¹⁴⁸ In addition, E2 has been shown to down-regulate cardiac Kv4.3 gene expression, an effect that has been linked to the longer QT interval observed in pregnancy. ¹⁴⁸

Interestingly, alterations in markers of pathological cardiac hypertrophy—such as α- and β-MHC (myosin heavy chain) switch or changes in atrial natriuretic peptide, phospholamban, or SERCA (sarco/endoplasmic reticulum Ca²⁺-ATPase), all of which have been implicated in the pathophysiology of RV failure in PAH^153,154—have not been detected in pregnancy. ¹⁴⁸ Additional studies that further characterized cardiac hypertrophy in late pregnancy in mice demonstrated increased cardiac angiogenesis, absence of fibrosis, and decreased expression of matrix metalloproteinase 2, 15, and 17. ¹⁵⁵ Despite an increase in cardiac contractility, LV pressures are increased, suggesting a component of cardiac dysfunction. ¹⁵⁵ These changes were reversed in the postpartum period. ¹⁵⁵ Furthermore, the myocardium of late pregnancy exhibits decreased proteolytic activity of the ubiquitin-proteasome system, as well as decreased oxidative stress. ¹⁵⁶ The decrease in proteasome activity appears to be independent of E2 levels. ¹⁵⁶

Sex hormone effects on cardiac function. A significant body of work has demonstrated protective effects of E2 and/or female sex in various forms of LV injury, all of which would expected to benefit RV function in pregnant PH patients. In particular, several studies in various types of LV injury (e.g., chronic pressure overload or myocardial ischemia) have demonstrated protective effects of E2 and/or female sex on cardiac function and mortality.^157–165 These improved outcomes are accompanied by favorable estrogen receptor–mediated E2 effects on myocardial metabolism, inflammation, microRNA signaling, fibrosis, matrix remodeling, and apoptosis.^157–166 E2 mediates these favorable adaptations to LV injury by targeting multiple cardiac cell types, such as cardiomyocytes, endothelial cells, fibrocytes, progenitor cells, and inflammatory cells (reviewed by Murphy ¹⁶⁷ ). Interestingly, E2 has consistently been shown to decrease cardiac hypertrophy in LV injury models and in in vitro studies,^{160,164,165,168–170} a finding that seems to contradict findings of E2-mediated increases in cardiac hypertrophy during pregnancy. ¹⁴⁸ Differences in the type and duration of cardiac hypertrophy (adaptive versus maladaptive) may explain this discrepancy, at least in part.

Despite a significant number of studies investigating effects of E2 and sex on LV function in pregnancy and disease, how pregnancy and sex hormones affect the function and molecular makeup of the normal RV remains relatively unexplored. Furthermore, the effects of sex hormones and pregnancy on an already diseased RV have not been well characterized. In particular, it is unknown whether and how pregnancy and sex hormones affect physiological or biochemical parameters in the RV once these are already altered by chronic pressure overload. What is currently known is that both healthy women and women with PAH exhibit RV function superior to that of their male counterparts.^{109–111,171} Furthermore, in female hormone-replacement therapy users, RV ejection fraction correlates with E2 levels, suggesting either a direct effect of E2 on RV function or a more indirect effect via modification of RV afterload. ¹¹⁰ Along those lines, basic-science investigations in nonpregnant rodents with PH demonstrated protective estrogen receptor–mediated E2 effects on RV function that are associated with decreased RV stress signaling, increased RV capillary density, and decreased fibrosis as well as with decreased expression of the matrix-remodeling enzymes ADAM15, ADAM17, and osteopontin.^112,115,172 Taken together, these data suggest salutary E2 effects on RV function in PH. Whether these animal findings translate to human disease is currently unknown.

Similarly, DHEA has been shown to exert RV-protective effects in rodent models of PAH, where its administration was associated with decreased oxidative stress, decreased apoptosis, and decreased activation of cardiac-remodeling mediators (e.g., Rho kinase, STAT3, and NFATc3), ¹⁷³ as well as with improved mitochondrial bioenergetics and antioxidant mechanisms. ¹⁷⁴ In humans, DHEA levels are associated with higher mass and higher RV SVs in healthy women. ¹¹⁰ Testosterone, on the other hand, is associated with lower RV ejection fraction and higher RV mass in healthy subjects, ¹¹⁰ and testosterone increases RV hypertrophy and fibrosis in male mice undergoing PA banding. ¹⁷⁵ Sex hormone effects on RV function in PAH are summarized in Figure 3.

Hormone-mediated cardioprotection and the peripartum period. While these studies suggest beneficial effects of major pregnancy hormones on RV function, they do not, however, explain the high prevalence of pregnancy-associated RV failure in PAH. One potential explanation is that despite potential beneficial effects of E2 and DHEA on RV function, increased testosterone levels during pregnancy^176,177 negatively affect RV performance. Furthermore, rapid decreases of E2 and DHEA levels after delivery may contribute to RV failure in the postpartum stage. In this scenario, the negative effects on RV function, occurring simultaneously with worsening pulmonary vasoconstriction, would provide a “perfect storm” and cause the RV to decompensate. Alternatively (or in addition), the RV stress associated with the hemodynamic changes and volume shifts during pregnancy, and particularly the peripartum period, may “overwhelm” E2- and DHEA-mediated protection and thus induce RV failure, especially in light of the known transient decrease in cardiac function that normally occurs during pregnancy.^54,55,156

In sum, pregnancy is associated with a number of adaptive changes in the cardiopulmonary compartment that serve to accommodate the increased CO and to meet the increased cardiovascular demands of the pregnant body. In PAH, however, several of these compensatory mechanisms are attenuated or simply overwhelmed. Pregnant PAH patients usually die of RV failure and cardiovascular collapse, with the majority of deaths occurring in the early postpartum period. The changes in hemodynamics and volume status observed during pregnancy appear to be a major cause of the worse outcomes in pregnant PAH patients. The role of sex hormones is not yet clear, as some studies show beneficial effects (e.g., vasodilation) while others suggest that they exert harmful effects (e.g., proliferation of PASMCs) on the pulmonary vasculature. Data for cardiac function are more clear-cut, as there is a solid body of evidence that supports beneficial effects of E2 and DHEA on cardiac function, even though studies specifically focusing on the RV are sparse and primarily in animals. In addition, studies investigating the effects of sex hormones specifically on RV function in the setting of pregnancy do not exist; this represents a significant knowledge gap in the field. The contribution of the abrupt fall in sex hormones after delivery to the high prevalence of cardiac decompensation in the postpartum period requires further investigation.

Effects of pregnancy on non–group 1 PH

The great bulk of current literature on sex hormone effects on the pulmonary vasculature and the potential effects of pregnancy are focused on PAH (WHO group 1 disease) or on hypoxia-associated PH (group 3). The effects of sex hormones on WHO group 2, 3, 4, and 5 PH and their potential effects on the pulmonary vasculature in pregnancy are essentially unstudied and are a major knowledge gap at present.

PPHN: potential maternal influences

Studies suggest a possible association between maternal use of the selective serotonin-reuptake inhibitor (SSRI) fluoxetine late in the third trimester of pregnancy and the risk of PPHN in the infant.^178,179 In further studies it was shown that an increased risk for PPHN was indicated for high maternal age, first parity, high maternal body mass index (BMI), and possibly maternal smoking. Adjusting for these variables and year of birth, an association between maternal use of SSRIs and PPHN in births after 34 completed weeks was still identified, with a risk ratio of 2.4 (95% confidence interval [95% CI]: 1.2–4.3) when based on women who reported the drug use in early pregnancy. When a subgroup of the women who had prescriptions for SSRIs from the antenatal care later in pregnancy were studied, the risk ratio estimate was 3.6 (95% CI: 1.2–8.3). ¹⁸⁰ However, these findings remain controversial, and some conclude that, because of sample sizes and quality issues in studies, the absolute risk cannot be determined. ¹⁸¹ If there is a risk, the mechanism behind the association between SSRIs and PPHN is unclear, but an increased risk for respiratory problems after maternal use of SSRIs is well known, and PPHN could be a rare component of this association. Possible etiological mechanisms are polymorphisms in the promoter alleles of the serotonin transporter gene ¹⁸² and in enzymes involved in catecholamine metabolism, ¹⁸³ which may predispose to neonatal adverse effects, including respiratory distress. ¹⁸⁴ Experimentally, it has been suggested that prolonged exposure to SSRIs can induce PPHN through direct effects on the fetal pulmonary circulation via serotonin. Serotonin causes pulmonary vasoconstriction, contributes to maintenance of high PVR in the normal fetus, and hence may contribute to the hypertensive effects of SSRIs.^185,186 A recent review concluded that the risk of PPHN is increased for infants exposed to SSRIs in late pregnancy, independent of the potential moderator variables examined. Although the statistical association was significant, clinically the absolute risk of PPHN remained low even in the context of late exposure to SSRIs. ¹⁸⁷

Summary

Summary of recommendations

Pregnancy outcomes in the preprostacyclin era

In 1996, epoprostenol was approved for treatment of PAH in the United States. Although not approved by regulatory agencies outside the United States, epoprostenol and its analog iloprost have also been used in Europe and other regions. Historical studies from the preprostacyclin treatment era, including systematic reviews of the published literature, identified a risk of maternal death in the range of 30%–56%. ¹ In a review of the published literature between January 1978 and December 1996, Weiss et al. ⁴ identified 73 patients with Eisenmenger's syndrome, 27 with idiopathic PAH (IPAH, formerly called “primary PH”), and 25 with PAH associated with other conditions (formerly called “secondary PH”). All patients, except for 1, delivered at 32 weeks or later. Pulmonary hypertensive crisis after oxytocin use occurred in 1 patient. Maternal mortality was 36% in Eisenmenger's syndrome, 30% in IPAH, and 56% in associated PAH. All fatalities, except for 3, occurred within 35 days after delivery. Late maternal death several months postpartum was reported in 3 patients with Eisenmenger's syndrome. The causes of death were pulmonary hypertensive crisis with therapy-resistant heart failure (n = 13), sudden death (n = 7), autopsy-confirmed pulmonary thromboembolism (n = 3), cerebral thromboembolism (n = 1), and rupture and dissection of the PA (n = 1). Previous pregnancies, timing of diagnosis and hospital admission, operative delivery, and diastolic PA pressure were significant univariate maternal risk factors. Late diagnosis (P = 0.002, odds ratio: 5.4) and late hospital admission (P = 0.01, odds ratio: 1.1 per week of pregnancy) were independent predictive risk factors of maternal mortality. Twenty-two neonates (weight: 2,372 ± 640 g) survived. One stillbirth and 2 neonatal deaths due to congenital anomalies resulted in a fetal/neonatal mortality rate of 12% (95% CI: 3%–31%). Of note, 1 parturient was treated with heparin and urokinase; operative delivery was assisted by extracorporeal circulation, and postpartum treatment with coumarin, nifedipine, prostaglandins, and diuretics was associated with the patient's survival. Nevertheless, in aggregate, these data suggest that pregnancy in the preprostacyclin era was associated with extremely high maternal and fetal morbidity and mortality.

Pregnancy outcome reports overlapping the pre- and postprostacyclin eras

Mortality and morbidity in pregnant PAH patients improved somewhat over the next years but remained exceedingly high. Bonnin et al. ²² reviewed the charts of 14 pregnant women with 15 pregnancies over a period of 10 years (1992–2002). The overall mortality in this study was 36% (5 of 14 patients). In this series, 3 cases of early deterioration occurred, at weeks 12, 20, and 24 of gestation, in only 1 of which did the patient survive, suggesting that early deterioration should lead to therapeutic abortion. Five of the seven other cases of deterioration occurred during the third trimester.

Experience from Japan was reported in 42 pregnant women who were studied retrospectively from January 1982 to December 2007. ¹⁹² Patients were divided into mild-PAH (systolic PAP of 30–50 mmHg on echocardiography or mPAP of 25–40 mmHg by catheterization; n = 14) and severe-PAH (systolic PAP ≥ 50 mmHg on echocardiography or mPAP ≥ 40 mmHg by catheterization; n = 28) groups. The cohort included mostly patients with IPAH or congenital heart disease–associated PAH. All patients with severe PAH before pregnancy had evidence of worsening PA pressures in their last trimester of pregnancy, while most of those with mild PAH had mild or no increase in PA pressures, although this latter group was assessed only by noninvasive methods. Patients with mild PAH delivered mostly at term, whereas those with severe PAH had earlier deliveries. The New York Heart Association (NYHA) functional class worsened in all but 2 patients as pregnancy progressed. There was 1 death immediately after intubation for delivery. Although the authors did not report the use of specific PAH therapies, they mention the use of iv epoprostenol in 3 cases.

Bédard el al. ²¹ conducted a comparative analysis of articles published between January 1997 and September 2007. Overall mortality was 25%. Parturients who received general anesthesia were 4 times more likely to die than parturients receiving regional anesthesia. Unlike the report from Weiss et al., ⁴ primigravidae were at higher risk of death than parturients with previous pregnancies. Seventy-eight percent of deaths occurred within 1 month postpartum (15 deaths postpartum vs. 3 deaths during pregnancy). Neonatal/fetal deaths occurred in 10%, 7%, and 13% of patients with IPAH, congenital disease–associated PAH, and patients with associated PAH, respectively.

Pregnancy outcomes in the era of PAH-specific therapies

Over the past decade, there have been several reports of patients with PAH successfully managed during pregnancy, suggesting an improved outcome with current therapies.^196–198 This literature, however, is subject to publication bias, may reflect the most successful management of PAH in pregnancy, ¹⁹⁹ and should be interpreted with caution.

In a retrospective series conducted between 1999 and 2009 at 5 US medical centers, experience from 18 pregnant PAH patients was reported (50% of patients had congenital heart disease). ¹⁹⁴ Six underwent pregnancy termination at a mean gestational age of 13 ± 1 weeks, with no maternal deaths or complications. Twelve elected to continue pregnancy. PAH-specific therapy was administered to 9 (75%) at the time of delivery, consisting of sildenafil, iv prostanoids, or combination therapy. All parturients underwent cesarean section at 34 weeks. There was 1 in-hospital death and 1 additional death 2 months postpartum, responsible for a maternal mortality rate of 16.7%.

Experience from China is evidenced in a retrospective review of 30 consecutive parturients with PAH hospitalized at Peking Union Medical College Hospital from January 1999 to December 2008. ²⁰⁰ The majority of patients were transferred to the PH center at near-term gestation; therefore, they were not followed in a center throughout their pregnancy. Eight patients had IPAH, 7 had congenital heart disease–associated PAH, 10 had rheumatic heart disease–associated PH, and 5 had PH due to other etiologies. PAH-specific therapies were used only in 2 patients with congenital heart disease–associated PAH. The condition of 57% deteriorated to NYHA classes 3 or 4 during pregnancy. Three patients were hospitalized, at 16, 18, and 23 weeks of gestation, for therapeutic abortion because of severe hemodynamic instability, and only 1 survived. The overall maternal mortality rate was 17%, but patients with Eisenmenger's syndrome had 50% mortality. There were 4 fetal/neonatal deaths (13%), and 16 infants were born preterm. All 26 liveborn infants survived.

Published Portuguese experience consists of 3 cases of IPAH from a single center. ²⁰¹ One patient had a therapeutic abortion, but her PAH deteriorated after abortion, with subsequent demise. The second patient (in whom PAH was well controlled before pregnancy) died 4 months after therapeutic abortion. A third patient (a calcium channel blocker long-term responder) underwent cesarean delivery and was managed intraoperatively with iv prostanoid. Although the patient survived pregnancy and delivery, she ended up requiring long-term PAH-specific therapies.

In 2012, Rosengarten et al. ²⁰² reported their recent experience from 7 patients with PAH who had a total of 9 pregnancies and for whom a careful multidisciplinary approach was implemented. All but 1 patient were treated with iv prostacyclins, and all underwent planned cesarean sections at term. Two patients died within 2 weeks after delivery (22% mortality), and there were no fetal deaths.

In a prospective study from 13 participating centers from Europe, the United States, and Australia, Jaïs and colleagues ²³ reported data from 26 pregnancies between July 1, 2007, and June 30, 2010, and their 3-year outcomes. During pregnancy, the patients were regularly seen at their PH centers, usually at 2–4-week intervals. Only one center in France performed right heart catheterizations as part of their regular assessment during pregnancy. A total of 16 pregnancies (62%) were considered successful, defined as survival of the mother and the baby without complications. Six patients underwent induced abortion, and 2 had a spontaneous abortion. Both women with spontaneous abortions died. Three women died in the early postpartum period as a result of right heart failure, whereas 1 required urgent heart-lung transplantation. The outcome of pregnancies was better in patients with lower PVR (500 ± 352 dyn-s/cm ⁵ ), whereas patients with a very high PVR (1,667 ± 209 dyn-s/cm ⁵ ) died or required transplantation. Of note, 8 (50%) of the 16 women who had successful pregnancies were so-called vasodilator responders and had nearly normal baseline hemodynamic parameters on calcium channel blocker therapy. The other women with successful pregnancies were calcium channel blocker nonresponders; however, most of them had well-controlled PAH while receiving PAH-specific therapy. Maternal outcome during the year after delivery was analyzed for the 16 women who had successful pregnancies. Two (13%) of them experienced clinical deterioration requiring intensification of PAH therapy. Neither was a long-term responder to calcium channel blockers.

The UK experience is reported in 2 papers. In the earlier report, Kiely et al. ¹⁹⁵ analyzed retrospectively the management of 10 pregnancies in 9 women who chose to continue with their pregnancies between 2002 and 2009 in a specialized PH center. All women were treated with 4–7 doses of nebulized iloprost per day, with 2 patients being transitioned to iv prostacyclin and 3 patients additionally receiving sildenafil. Nine patients underwent planned cesarean section. All women received regional anesthesia and were monitored during the peripartum period in a critical care setting. All women survived the pregnancy and postpartum period, and all infants were free from congenital abnormalities. One woman died 4 weeks after delivery, following patient-initiated discontinuation of therapy. Long-term follow-up in the remainder of the patients was a median of 3.2 years (range: 0.8–6.5 years). A second UK report describes retrospectively the outcome of 12 pregnancies in 9 women with PAH between 1995 and 2010 at one center. ¹⁹³ All women delivered by cesarean section (7 elective and 2 emergency deliveries) under general anesthesia, except for 1 emergency and 1 elective cesarean performed under regional block. There were 2 maternal deaths (in 1995 and 1998), 1 related to preeclampsia and 1 to arrhythmia. Maternal morbidity included postpartum hemorrhage (5 cases) and 1 postcesarean evacuation of a wound hematoma. There were 9 live births, 3 first-trimester miscarriages, and no perinatal deaths.

A most recent prospective international registry including centers from the United States and Europe reported on 26 pregnant PAH patients (17 with IPAH, 9 with associated PAH). Six chose elective abortion, 16 delivered a healthy baby and survived without transplantation, and 4 died or underwent urgent transplantation. ²³ Of the 16 patients who successfully delivered, 2 experienced clinical deterioration requiring intensification of PAH therapy.

Taken together, these data suggest that outcomes of pregnant patients have improved with the availability of new PAH therapies, advances in surgical and perioperative management, and use of a team-based, multidisciplinary approach. However, publication bias may exist, and despite the availability of new drugs and technologies, fetal as well as maternal morbidity and mortality remain high.

Summary of recommendations

Barrier methods and other methods

Barrier methods of contraception include diaphragms, cervical caps, and sponges. Because barrier methods are event driven and often have a higher failure rate when used alone, they are not recommended as sole forms of contraception in patients with PH.

Fertility-awareness methods, such as “awareness-based” avoidance of intercourse and lactation-induced amenorrhea, are similarly ineffective^205,213 and are not recommended to prevent pregnancy in PH patients. Spermicides are among the least effective methods of contraception and are thus not recommended to be used alone as contraception in PH. ²¹⁴

Permanent sterilization

Because pregnancy will be dangerous for patients with PH for the foreseeable future and because there is no perfect reversible form of birth control for these women, some centers favor permanent contraception. Options include tubal ligation or a device implanted into the fallopian tube (Essure device, Conceptus, Mountain View, CA). Male sterilization (vasectomy) can be combined with a second form of female birth control, alternatively.

Tubal ligation can be performed laparoscopically or via a minilaparotomy. Both procedures are performed with the patient under general anesthesia with positive-pressure ventilation. Although commonly used in healthy women, a laparoscopic approach entails greater risk for the PH patient than for the general population. Patients are positioned with the head down to shift abdominal contents away from the pelvis, resulting in diaphragmatic impingement of the thoracic cavity. Furthermore, manipulation of the cervix is frequently associated with a vasovagal response, and insufflation of carbon dioxide gas into the abdomen causes increased abdominal pressure and may result in a paradoxical CO₂ embolism. For these reasons, a minilaparotomy may be a safer approach. Although still requiring general anesthesia, a minilaparotomy is performed in the supine position, and CO₂ insufflation is not needed.

Where available, hysteroscopic sterilization is the preferred option for permanent sterilization. This procedure entails the hysteroscopic placement of nickel-titanium coils wound around polyethylene fibers that are inserted into the fallopian tubes. No incision is required. Patients require minimal or no anesthesia, as frequently conscious sedation is adequate to perform the procedure. After placement, the device stimulates an inflammatory response that results in tubal occlusion within 10–12 weeks. The procedure is generally safe and effective, even in high-risk patients. In a retrospective cohort study of hysteroscopic sterilization, 18 women with severe cardiac disease, including 5 with PH, were compared with 157 women without cardiac disease. ²¹⁵ No intraoperative complications were reported, and follow-up tubal-occlusion rates were 100% in the cardiac group, compared with 98% in the control group. A report of 4 patients with PAH described similar outcomes: no procedure or anesthesia-related complications were encountered. ²¹⁶

Emergency contraception

In cases where emergency contraception is required, hormonal methods may be considered, with the same caveats discussed above. Alternatively, placement of a copper IUD has also been reported to be a successful method of emergency contraception, with the added benefits of providing a lasting method of contraception going forward and avoidance of exposure to high-dose sex hormones. For patients using the endothelin receptor antagonist bosentan, the dose of emergency contraception will have to be increased because of the presence of drug-drug interactions.

Summary of recommendations

Genetic counseling in IPAH and HPAH

Mutations in BMPR2 can be found in about 80% of families affected by HPAH.^8,2 Other genetic mutations have recently been identified in HPAH, including caveolin 1 (CAV1), ²²⁰ KCNK3, ²²¹ and EIF2AK4, ²²² that further explain a proportion of the remaining 20% of HPAH. Extensive research in the past decade has focused on the epidemiology of HPAH and downstream signaling associated with BMPR2 mutation, but there is no cure at present for this genetic mutation, and patients with HPAH are treated similarly to patients with IPAH. The BMPR2 mutation is a dominantnegative gene with only weak penetrance, such that approximately only 20% of people with the mutation will develop PAH while the remaining 80% do not develop clinical PAH.^218,219

In addition, a small percentage of patients with IPAH also have a BMPR2 mutation but do not have any family members affected by PAH, because of either the low penetrance or de novo mutations. ²²³ These IPAH patients with BMPR2 mutation can, however, transmit the mutation to their biologic offspring, who are at risk to develop PAH if they receive this allele. The frequencies of the other genes associated with HPAH in the IPAH population are unknown at present. Germ-line genetic mutations have not been described in non–group 1 PH and in PAH that is not idiopathic or heritable.

Clinically available genetic testing can detect most BMPR2 mutations, and some facilities detect CAV1, ALK1, and KCNK5. Others, such as EIF2AK4 and other newly discovered mutations, are screened for only in the research setting at present.

Given the complexity of genetics in HPAH and IPAH, patients wishing to become pregnant should be offered genetic counseling before conception. In the process of genetic counseling, patients will be aided in understanding the risks and benefits of genetic testing and will then be equipped to make an informed decision regarding whether or not they want to undergo genetic testing. They may also consider forms of preimplantation genetic testing or postconception testing for the fetus if they decide to pursue pregnancy. Given the potential psychological and financial risk associated with genetic testing, it is not recommended to undergo genetic testing in the absence of genetic counseling.

Genetic counseling in patients with PAH other than idiopathic or heritable and non–group 1 PH.

Patients with other forms of PAH or non–group 1 PH are not recommended to undergo genetic counseling or testing for BMPR2 mutation unless there are unusual circumstances.

Summary of recommendations

Therapeutic abortion

Because of the excessive risk to the mother and fetus discussed in “Basic biology of PH and pregnancy,” when pregnancy occurs in PH patients, termination should be offered regardless of WHO FC or other markers of prognosis. ²²⁵ Discussion of therapeutic abortion, using a shared-decision model, has been proposed to establish a common set of goals and optimize maternal medical care. ²²⁵ The first trimester is the safest time for elective pregnancy termination; however, in the PH patient, pregnancy termination carries greater risk than in the general population and should be performed in an experienced center.^224,226 Uterine dilatation and evacuation is the safest procedure. ²²⁷ If surgical evacuation is not feasible, medical abortion using prostaglandins E₁ or E₂ or misoprostol, a synthetic prostaglandin structurally related to prostaglandin E₁, can be administered to evacuate the uterus. ²²⁷ These drugs are absorbed into the systemic circulation and can lower SVR and blood pressure and increase heart rate; these effects are greater with prostaglandin E₂ than with E₁. ²²⁷

While therapeutic abortion is most commonly performed during the first trimester in the pregnant PH patient, it should also be considered in the second trimester up to the point of fetal viability. After that point, early delivery may be considered if clinically indicated. Different cultural and/or religious preferences and legal obligations are present in the diverse countries of pregnant PH patients. These factors may influence timing, acceptability, and options for therapeutic abortion.

First trimester

Despite strong recommendations for therapeutic abortion, some pregnant PH patients may decide to continue pregnancy. In this case, management should be provided by a multidisciplinary team in a center experienced in the treatment of not just PH patients but also pregnant PH patients. At the minimum, the team should include a PH specialist, a cardiologist, an obstetrician, an anesthetist specialized in managing high-risk pregnancies, and a neonatologist.^{195,224–226} Pharmacists, intensivists, and social workers are frequently needed as well. In addition, in the case of associated forms of PAH, specialists in the underlying disease should be involved. Patient-specific multidisciplinary management plans should be put in place. ¹⁹⁵

A careful baseline assessment of pregnancy risk during this trimester should include history and physical examination, echocardiography, brain natriuretic peptide measurement, and 6-minute-walk testing. A discussion of PAH-directed therapy in pregnancy follows in “Role for PAH-directed therapy in pregnancy and delivery.” Patients should be followed closely, typically monthly, but assessment frequency should be tailored to individual patients and their risks. During the first trimester, the volume changes associated with pregnancy are less pronounced than those in future trimesters; thus, edema and swelling symptoms should trigger an evaluation of PH status. As discussed above in “Physiologic changes of pregnancy/concerns in PH,” hormonal changes associated with early pregnancy may alter pulmonary vascular disease and precipitate worsening. Even in the absence of symptoms, regular assessment of PH status is advisable, using, in addition to physical examination, echocardiography, brain natriuretic peptide, 6-minute-walk testing, or a combination of all at regular intervals during this trimester. Management of PAH-specific therapy is discussed in “Genetic counseling in PH.”

Second trimester

As discussed in “Physiologic changes of pregnancy/concerns in PH,” a major change in the second trimester is volume expansion and increase in CO, which is achieved in part through vasodilation and recruitment of the pulmonary vascular bed. In most forms of precapillary PH (WHO groups 1, 3, 4, and 5), the diseased pulmonary vasculature cannot accommodate the increased CO. Patients with group 2 PH may also have difficulty accommodating increased flow though the pulmonary vasculature. As a consequence, RV volume loading and subsequent failure may occur. A dilated RV can, in turn, reduce left heart filling, thus further compromising CO. As a function of these volume demands, decompensation often occurs in the second trimester, corresponding to peak blood volume and CO. Vigilant follow-up though the previously discussed multidisciplinary team is essential, and frequency of visits may have to be increased to every 2 weeks, for example. As in the first trimester, follow-up evaluation should include history and physical examination, 6-minute-walk testing, and brain natriuretic peptide measurement. Echocardiography frequency may be increased to every 2 weeks if concerns arise during this trimester of worsening right heart function.

In patients who have developed right heart failure symptoms, standard therapy including diuretics, oxygen therapy if indicated, and, for patients with severe symptoms or failure of outpatient therapy, hospitalization may be considered. In patients requiring diuretics, furosemide is preferred, as spironolactone has antiandrogenic properties and is pregnancy category D. As with all right heart failure, judicious fluid restriction, low-salt diet, and oxygen supplementation are indicated where appropriate. To avoid caval vein compression, the patient should be advised to lie in the lateral position. Management of PAH-specific therapy is discussed below in “Role for PAH-directed therapy in pregnancy and delivery.”

Third trimester

As the major volume expansion has occurred in the second trimester, the third trimester's main concerns are preparation for delivery and ensuring continued safety of the mother and fetus. In the third trimester, weekly follow-up is suggested (Fig. 4). In this trimester, echocardiograms are recommended every 1–2 weeks to monitor the cardiac status and adjust PH medications (Fig. 4). As in the second trimester, important nonpharmacologic therapy involves limiting sodium and fluid intake during pregnancy. Limiting daily fluids to 1.5–2 L, restricting sodium intake, and administering diuretics in the third trimester can reduce RV overdistention and avoid cardiac decompensation. Elective cesarean section is advisable when the fetus becomes viable. A multidisciplinary approach has been reported to lead to a favorable maternal-fetal outcome. ¹⁹⁵ Close communication between high-risk obstetrician, anesthesiologist, and PH physician help in the proper and timely management of these high-risk patients. ¹⁹⁵

OPEN IN VIEWER

Figure 4.

Recommended evaluation of and follow-up for a pregnant patient with pulmonary arterial hypertension (PAH). BNP: brain natriuretic peptide; ETRA: endothelin receptor antagonist; FC: World Health Organization function class; ICU: intensive care unit; LMWH: low-molecular-weight heparin; PH: pulmonary hypertension; RV: right ventricular; 6MWT: 6-minute walk test.

Delivery

The mode of delivery should be carefully selected in the pregnant PH patient, as it may influence the outcome. Cesarean section is the preferred mode of delivery. ²²⁴ Although vaginal delivery is usually associated with fewer bleeding complications and infections in the healthy population, ²²⁹ the hemodynamic and physiological changes associated may be detrimental to the mother with PH. The advantages and disadvantages of different modes of delivery are discussed below.

Monitoring and general considerations during delivery

During labor and delivery, continuous monitoring of electrocardiogram, pulse oximetry, central venous pressure, and intra-arterial blood pressure should be routine. ²²⁴ While PA catheterization may be of benefit in selected cases, there is no consensus regarding its routine use during delivery. ²¹ Several centers manage patients with central lines and close monitoring of RV function (e.g., by echocardiography). The use of noninvasive hemodynamic monitoring in this setting may be of benefit but has not yet been validated. Close attention must be paid to avoiding conditions that may lead to PA vasoconstriction and worsening RV function; these include hypoxia, hypercarbia, and acidosis. Hemodynamically relevant arrhythmias should be aggressively corrected. Anxiety- and pain-induced catecholamine release may negatively affect hemodynamics; the use of different anesthetic methods is discussed in detail below. In order to optimize oxygen delivery as much as possible, significant anemia should be corrected. The patient should be as euvolemic as possible, and major fluid shifts must be avoided as much as possible. Vasopressors and inotropes should be readily available for hemodynamic support. Intravenous prostacyclins should be readily available; however, the role and management of PAH-specific medications during delivery is discussed in detail in “Role for PAH-directed therapy in pregnancy and delivery.” Individualized management plans for each patient must be discussed and updated before delivery.

Vaginal delivery

This mode of delivery is usually avoided in PAH patients, as it carries a multitude of disadvantages. First, vaginal delivery is associated with frequent use of the Valsalva maneuver, which increases intrathoracic pressure and thus decreases venous return. As the CO in PH patients requires adequate preload, the Valsalva maneuver may critically and rapidly decrease preload and thus precipitate cardiopulmonary collapse. Second, labor-induced vasovagal responses can decrease the venous return and lead to cardiopulmonary collapse. Third, the pain of childbirth may lead to sympathetic nervous system stimulation. As a result, there may be an increase in heart rate and increases in systemic and pulmonary vascular tone, with resultant hemodynamic instability. Labor-induced acidosis, hypercapnea, or hypoxia may further contribute to increases in PA pressures. After delivery of the fetus, an increase in maternal blood volume occurs because of autotransfusion of blood from the contracting uterus and from shifting of peripheral edema from the extravascular compartment into the systemic vasculature. In addition, venous return may increase from the relief of inferior vena cava obstruction by the gravid uterus. Finally, agents used to induce labor may precipitate clinical deterioration. Low-dose oxytocin can be used with caution to induce labor, as it can increase PVR. Because of the potential for hemodynamic instability due to the aforementioned factors, vasopressor or inotropic support may be required. If vaginal delivery is chosen, effective analgesia is imperative. Nitrous oxide should be avoided because of its potential to vasoconstrict the pulmonary vasculature, ²³⁰ and epidural analgesia should be considered.

Cesarean section. Cesarean section is the preferred mode of delivery and should be used unless not available or in cases of emergencies.^22,195 Cesarean section bypasses the hemodynamic complications associated with labor discussed above and also the autotransfusion associated with vaginal contractions. Elective cesarean section avoids labor and allows for careful, multidisciplinary planning and preparation of anesthesia, optimization of hemodynamics, and development of contingency plans. Moreover, cesarean section does not appear to negatively influence patient outcomes. ¹⁹⁷ The choice of anesthesia is important when cesarean section is being considered and should be preplanned during early multidisciplinary meetings with the anesthesiologist, who should be aware of the hemodynamic changes associated with PH (often an obstetric or cardiac anesthesiologist), and with an experienced maternal-fetal medicine specialist. The aim of the anesthesiologist is to achieve an optimum balance between anesthesia and side effects. Most obstetric units would consider regional anesthesia in preference to general anesthesia for this procedure. This will reduce risks associated with general anesthesia (e.g., endotracheal intubation and positive-pressure ventilation, which may increase PVR and RV afterload). However, a regional approach is not always possible, and plans should be made to proceed with general anesthesia if necessary.

Timing of delivery. The optimal timing of planned elective delivery in women with PAH is uncertain and involves an assessment of the maternal risk of continuing pregnancy and risk of neonatal preterm delivery. In stable women, planned delivery around weeks 34–36 is recommended, with delivery before this if there is evidence of symptomatic decline.

Anesthesia in pregnant PH patients

General anesthesia. General anesthesia is known to depress cardiac contractility, increase PVR via positive-pressure ventilation, and increase PAP during laryngoscopy and intubation. ²³¹ Studies by Bédard and others ²¹ suggest that parturients who receive general anesthesia were more likely to die than parturients receiving regional anesthesia. However, caution should be shown when interpreting these results, as selection bias may be shown toward the sicker parturient receiving general anesthesia. If general anesthesia is required, the patient should be ventilated with the lowest airway pressure and tidal volume possible. A lung-protective ventilation strategy similar to what is used in patients with acute respiratory distress syndrome has been advocated. ²³² However, care must be taken to avoid hypercarbia. Catecholamines and iv prostacyclins should be readily available for these patients, and the administration of catecholamines before anesthesia induction should be considered.

Epidural or combined spinal-epidural anesthesia. Epidural anesthesia with incremental doses has traditionally been considered the best approach to regional anesthesia. However, with the advent of modern equipment and techniques, low-dose combined spinal-epidural anesthesia is being increasingly used, because it provides a denser perineal sensory block than epidural anesthesia alone, with little additional risk of hypotension. ²³³

Spinal anesthesia. Single-dose spinal anesthesia should be avoided in this patient population because of the risk of a rapid rise in block height that may cause uncontrollable hemodynamic instability as sympathetic block rises and systemic hypotension occurs. ²²

Postpartum and long-term effects

In the hours and days after childbirth, blood volume is increased by autotransfusion of blood from the contracting uterus and by shifting of peripheral edema from the extravascular compartment into the systemic vasculature. In healthy subjects, these physiologic changes are normally well tolerated, but in the presence of severe pulmonary vascular disease, hemodynamic instability may develop, leading to cardiorespiratory failure and/or sudden death.^194,224 Parturition and the first postpartum week have been recognized as particularly vulnerable periods for patients with PAH.^23,194 Two systematic literature reviews have described the outcome of pregnancy in women with PH, covering a total of almost 30 years and comprising 198 pregnancies. The mortality of women with Eisenmenger's syndrome (N = 102) was 36% in the first review (1978–1996) ⁵ and 28% in the second review (1997–2007) ²¹ . Most of these women died in the first month after delivery, and the main causes of death were heart failure and sudden death, while pulmonary thromboembolism was another frequent cause. In women with IPAH (N = 56), mortality was 30% in the first ⁴ and 17% in the second ²¹ review. Similarly, nearly all fatalities in these patients occurred early in the postpartum period, and death was mainly due to heart failure, followed by sudden death and thromboembolism. Thus, the immediate postpartum period is a critical period for acute decompensation in PH patients. Patients should be closely monitored for several days postpartum; monitoring in an intensive care unit in the first few days after delivery is recommended. This is especially pertinent to high-risk patients, particularly those with right heart failure or evidence of hemodyanmic instability during delivery. Prophylactic anticoagulation is important in this period (see “Role for anticoagulation in the pregnant PH patient”). ²³⁴ As the volume expansion that occurs in pregnancy is sustained for up to 24 weeks postpartum, improvement in functional status after delivery may take several months.

Detrimental effects of vagal reactions in pregnant patients with PH

Vagal reactions and syncope can have catastrophic consequences in PH patients independent of the presence or absence of pregnancy, because they may be associated with a profound drop in CO. ²³⁵ However, in pregnant PH patients, the presence of a vagal reaction is particularly concerning, as the RV is extremely preload dependent and compensatory mechanisms are even more limited than in nonpregnant PH patients. It is therefore of utmost importance that potential inducers of vagal stimuli be recognized and avoided.

Vasovagal syncope is characterized by a temporary reduction of blood flow to the brain that is caused by sudden hypotension, change in heart rate, and/or change in blood volume or its distribution. ²³⁶ In general, because of their vasodilatory state, pregnant women are more prone to vasovagal syncope than their nonpregnant counterparts. ²³⁷ Since PH patients are more prone to syncope than healthy individuals, ²³⁵ the combination of pregnancy and PH therefore creates a “perfect storm,” generating a condition where (1) vasovagal syncope is more prevalent and (2) its consequences are particularly dangerous. Common inducers of vagal reactions during pregnancy include vena cava compression, rapid positional changes (especially getting up too fast), pain, anxiety, straining while pushing during vaginal delivery, hypovolemia, anemia, hyperthermia, and hyperventilation. Iatrogenic causes include administration of vasodilators, interventions that induce vasodilation and decrease venous return (e.g., induction of general or spinal anesthesia), manipulation of the cervix, and any procedure that causes pain and/or anxiety. Common preceding symptoms of vasovagal syncope include nausea, diaphoresis, lightheadedness, and pallor.

Because of the potential catastrophic consequences of vasovagal syncope in the pregnant PH patient, it is of utmost importance that triggers of such events be avoided as much as possible. If the trigger is unavoidable or has already occurred, rapid and focused intervention is critical. Counseling of the pregnant PH patient on inducers and symptoms of vasovagal syncope, as well as on appropriate interventions, is of highest importance. Similarly, medical professionals need to be aware of interventions known to raise vasovagal tone, and they need to be educated that PH patients are particularly susceptible to such reactions. If patients undergo any procedure that may induce vagal reactions, such as placement of an IUD or therapeutic abortion, physicians should place 2 iv lines and have atropine in the procedure room ready to administer.

Summary of recommendations

Prostaglandins

Prostaglandins are potent pulmonary vasodilators that may also act to enhance RV function.^238,239 Currently available parenteral prostaglandins—epoprostenol, treprostinil, and iloprost—are not known to be teratogens (Table 5) and are listed as pregnancy category B (epoprostenol and treprostinil) or C (iloprost).

OPEN IN VIEWER

Table 5.

FDA-assigned risk category for PAH medications

Drug	Pregnancy risk category a
Epoprostenol	B
Treprostinil	B
Sildenafil	B
Tadalafil	B
Nitric oxide	C
Iloprost	C
Bosentan	X
Ambrisentan	X
Macitentan	X
Riociguat	X

Note: FDA: US Food and Drug Administration; PAH: pulmonary arterial hypertension.

a

B: animal reproduction studies have failed to demonstrate a risk to the fetus, and there are no adequate, well-controlled studies in pregnant women; C: animal reproduction studies have shown an adverse effect on the fetus, and there are no adequate, well-controlled studies in humans, but potential benefits may warrant use of the drug in pregnant women despite potential risks; X: studies in animals or humans have demonstrated fetal abnormalities, and/or there is positive evidence of human fetal risk based on adverse-reaction data from investigational or marketing experience, and the risks involved in use of the drug in pregnant women clearly outweigh potential benefits.

They may thus be continued in patients who receive this class of medication before pregnancy, if treating physicians and patients agree that the benefits outweigh the relatively unknown risks. Prepregnancy use of prostaglandins, however, may be an indicator of a patient with more severe disease. In addition to the usual indications to begin prostaglandin therapy, such as WHO FC III or IV symptoms and reduced RV function,^14,240,241 other considerations in pregnancy may compel therapy with parenteral prostaglandins. Examples are worsening symptoms or deteriorating RV function despite therapy with phosphodiesterase inhibitors and nonparenteral prostanoids. It should be recognized, however, that even with parenteral prostanoids, maternal mortality may still be significant. ²⁰² In many centers, iv prostanoids are managed exclusively by the PH specialist; this includes accompanying the patient to the operating room.

In general, parenteral prostaglandins are recommended for pregnant patients with PAH in WHO FC IV or where there is evidence of severe RV impairment. In selected patients in WHO FC III followed in specialized centers, a trial of therapy with nebulized iloprost and addition of sildenafil has been successfully reported. ¹⁹⁵ When delivered by the parenteral route, gradual up-titration can be performed as per the usual protocol. Inhaled iloprost or treprostinil may be used in patients with less severe symptoms, and their safe and successful use in pregnancy has been reported.^195,198 However, in a patient experiencing deterioration while receiving inhaled-prostaglandin therapy, rapid conversion to parenteral forms (some publications have advocated for iv epoprostenol ²⁴² ) should be considered. If there is no rapid clinical improvement, immediate delivery should be considered because of the high risk of maternal death. In addition, inhaled epoprostenol or nitric oxide ²⁴³ has been administered continuously to acutely ill patients as a bridge to parenteral treatment initiation.

Phosphodiesterase 5 inhibitors

Both sildenafil and tadalafil are pregnancy category B (Table 5). There are several reports of successful pregnancies managed with sildenafil either alone or in combination with a prostaglandin.^{23,193,194,244,245} The use of tadalafil in pregnancy has not been reported. In larger case series of successful pregnancies, sildenafil has generally been used in combination with a prostaglandin.^23,194,195 As there is no published experience with tadalafil, we recommend a preference for sildenafil when phosphodiesterase inhibitors are used in pregnancy. Their use, however, should be restricted to either combination therapy with a prostaglandin or, if as a monotherapy, restricted to patients who have normal RV function and near-normal pulmonary hemodynamics or who refuse prostaglandin therapy. If phosphodiesterase inhibitor monotherapy is used, close follow-up for deterioration is indicated.

Calcium channel blockers

PAH patients who have an acute response to inhaled vasodilators such as nitric oxide (defined by a drop in mPAP by 10 mmHg to 40 mmHg or less with preserved CO) are candidates for treatment with calcium channel blocker therapy and have a substantially improved prognosis, compared to patients without an acute vasodilator response. ²⁴⁶ Durable response to calcium channel blocker therapy during pregnancy has been described. ²³ In fact, among the patients included in the only prospective registry of pregnant PAH patients, 8 of 16 patients who successfully delivered a baby and survived without transplantation were acute vasodilator responders. ²³ Calcium channel blockers are considered safe in pregnancy. ²⁴⁷ For those patients meeting strict criteria for an acute response to inhaled nitric oxide, we recommend therapy with calcium channel antagonists^14,246 and close follow-up, as discussed above. In pregnant PAH patients not meeting these strict criteria, there is no role for calcium channel antagonists. In this setting, calcium channel blockers may be detrimental because of their nonselective vasodilatory effects, with a concomitant drop in SVR, which, in the presence of an inadequate CO, may lead to circulatory collapse.

Endothelin receptor antagonists

Endothelin receptor antagonists (ambrisentan, bosentan, macitentan, and sitaxentan) are pregnancy category X (Table 5) and should be discontinued when a pregnancy is discovered in a patient with PAH, because of the known teratogenicity of these medications.

Soluble guanylate cyclase stimulator

Riociguat is pregnancy category X and should be discontinued when a pregnancy is discovered in a PAH or chronic thromboembolic PH (CTEPH) patient.

Summary for PAH-directed therapy during pregnancy

The advent of new therapies for PAH provides the clinician and the patient with a number of choices during pregnancy. These include the timing of treatment initiation in treatment-naive patients, the type of drug to be used, and the mode of drug delivery. Varied approaches in PH centers experienced in the management of PH in pregnancy reflect individual center experience, and where clear evidence of the benefit of one approach over another is lacking, pragmatic approaches will reflect local experience. ¹⁹⁵ It should be recognized that, despite improved outcomes and potential theoretical benefits of various treatment approaches, the medical treatment of PAH patients should be carefully chosen and individualized to take into account the severity of illness, comorbidities, social and cultural aspects, patient's wishes, and long-term goals.

Special considerations at delivery

The approach to the use of PAH therapies around the time of delivery will be influenced by the therapies that the patient has been taking during her pregnancy, her stability before planned delivery, and her progress in the peripartum period. When titrating these therapies, one needs to appreciate the physiological changes that are normally occurring in pregnant women and to be receptive to possible worsening of the pulmonary vascular disease and the effects of drugs used in the peripartum period. Oxytocic drugs, for example, can induce hypotension and tachycardia and should always be given by slow infusion.

Key to allowing titration of these therapies is careful monitoring of patients during the peripartum period in a high-acuity environment. Insertion of a central venous catheter to allow measurement of central venous pressure and sampling of blood to assess central venous saturation and insertion of an arterial line for measurement of blood pressure and blood gas sampling are highly recommended. ²⁴⁸ Swan-Ganz catheterization is not routinely used because of the risk of complications, but it can be considered in highly selected cases. Noninvasive measures of CO, such as lithium dilution, can be used to titrate treatment in patients with PAH.^195,248,249 This can allow a continuous measure of CO in addition to displaying systemic pressures, heart rate, and other core variables. However, its use has not been validated in prospective studies.

For stable patients receiving iv prostanoids, doses would usually remain unchanged before and during delivery. For patients not receiving parenteral prostanoids, a low-dose infusion of iv prostanoid started several hours before a planned cesarean section would allow the physician to assess the hemodynamic effects of this infusion and avoid problems associated with initiating prostanoids in the immediate peripartum period, when, in an unstable patient, it may be difficult to discern side effects of the drug from pulmonary hemodynamic instability. In these cases, a Swan-Ganz catheter may help assess the therapeutic effects of infused prostacyclins. If patients are receiving oral sildenafil, consideration should be given to using the iv preparation of the drug, particularly if there are any concerns regarding absorption, which can occur postoperatively, for example, following the development of ileus after cesarean section. For surgery under regional anesthesia, patients receiving nebulized iloprost or treprostinil therapy may continue to nebulize during the procedure, in addition to receiving a low-dose iv infusion during the peripartum period. Alternatively, nebulized epoprostenol may be substituted for intermittent inhalation of iloprost or treprostinil.

The treatment goals during delivery are to maintain systemic and right atrial pressures, to monitor fluid balance, and to avoid volume overload, particularly in the first 48 hours. A rising right atrial pressure after delivery may simply reflect fluid overload and can be managed by judicious use of diuretic therapy. If a patient's condition deteriorates with falling central venous saturation, rising right atrial pressure, and other features suggesting worsening of pulmonary vascular status, then increasing the dose beyond the background dose of iv prostanoid is usually appropriate, sometimes coupled with inotropic therapy, such as low-dose dobutamine. For patients developing hypotension, a cause should be sought, such as bleeding, sepsis, or low-output heart failure due to increasing afterload. In this setting, once volume status has been optimized, drugs such as vasopressin or noradrenaline should be considered. ²⁵⁰

While the treatment recommendations above assume access to PAH therapy, in locations where the full complement of PAH-directed therapy is not available, supportive care should be instituted. Such care is discussed in “Pregnancy management” above.

Summary of recommendations

Summary of recommendations