Hypertension in Pregnancy: Diagnosis, Blood Pressure Goals, and Pharmacotherapy: A Scientific Statement From the American Heart Association

Systemic vascular resistance decreases while plasma volume and cardiac output increase during pregnancy. There is a physiological drop in BP, often detectable before the end of the first trimester,^48,49 attributable to vasodilation.⁵⁰ Meta-analyses and high-quality longitudinal studies found that compared with BP at 10 or 12 weeks, the BP drop during the second trimester was on average 1 to 2 mm Hg.^51–53 There is wide interindividual variability, and BP trajectories likely relate to preexisting maternal health factors⁵³ such as chronic hypertension and require further clarification. Renal blood flow and glomerular filtration rate increase by 50% in normal pregnancy but are ≈30% lower in women with preeclampsia as a result of both decreases in renal blood flow and the ultrafiltration coefficient, attributable to endotheliosis in the glomerular capillary bed.⁵⁴ Plasma volume increases in normal pregnancy, and earlier studies have suggested that it may be decreased in women with preeclampsia.⁵⁵ However, multiple longitudinal and cross-sectional studies in preeclamptic women have demonstrated that suppressed plasma renin activity, high BP, decreased glomerular filtration rate, and frequent development of edema are more consistent with an overfilled, vasoconstricted circulation rather than true hypovolemia and underfilling.⁵⁶ Cardiometabolic changes in normal pregnancy are more pronounced in women who develop preeclampsia and include increased insulin resistance, total cholesterol, triglycerides, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol.⁵⁷ Hypercoagulability, a feature of normal pregnancy, may be exaggerated in preeclampsia and is caused by increased thrombin generation, fibrinogen, and activated protein C resistance and reduced protein S and fibrinolysis.⁵⁸

The diameter of the uterine spiral arteries increases greatly during normal pregnancy as a result of remodeling of the endothelium and vascular smooth muscle, stimulated by release of proteases from endovascular trophoblast and uterine natural killer cells.⁵⁹ Failure of spiral artery remodeling (ie, retention of smooth muscle) is a feature of preeclampsia^60,61 and leads to decreased utero-placental perfusion, demonstrated by noninvasive blood flow and perfusion studies using Doppler ultrasound or magnetic resonance imaging (Figure).⁶¹

Placental pathology attributable to rheological consequences includes villous architectural changes caused by turbulent jets entering the intervillous space at rates of 1 to 2 m/s (10–20 times normal), causing the rupture of anchoring villi and the formation of echogenic cystic lesions that are visible by ultrasound.⁶⁴ In addition, retention of vascular smooth muscle preserves the ability of spontaneous vasoconstriction and ischemia-reperfusion injury, which may result in oxidative stress.

Alterations in angiogenic factors are recognized as a likely consequence of abnormal placentation occurring in early pregnancy. Increased circulating soluble fms-like tyrosine kinase 1, an antiangiogenic factor of placental origin, leads to neutralization and decrease of proangiogenic factors such as placental growth factor and vascular endothelial growth factor, which then contribute to the hypertension and glomerulopathy characteristic of the maternal syndrome.⁶¹ Measurements of angiogenic biomarkers have been incorporated into risk stratification in several innovative therapeutic trials for preeclampsia prevention^65,66 but are not routinely used to guide clinical care in most countries, including the United States. An increased soluble fms-like tyrosine kinase 1/placental growth factor ratio may be particularly pronounced in women with early (<34 gestational weeks), severe preeclampsia, which has been designated by some as placental preeclampsia⁶⁷ because of the association between placental ischemia and adverse fetal outcomes (fetal growth restriction in particular). Preeclampsia occurring later in pregnancy, labeled maternal preeclampsia by some, has been associated with more pronounced maternal vascular dysfunction before pregnancy (secondary to hypertension, diabetes, or obesity), less pronounced placental pathology, and fewer fetal complications. In maternal preeclampsia, pregnancy acts as a physiological stress test that exacerbates preexisting endothelial dysfunction. This underscores the heterogeneity of HDP, whereby the extremes of clinical subtypes (early versus late, mild versus severe, and presence or absence of fetal growth restriction) may reflect distinct underlying mechanisms.⁶⁷ Sharp discrimination between maternal and placental preeclampsia is overly simplistic and artificial because both processes likely play a role but with varying contributions. Regardless of the clinical subtype, diagnosis and treatment of hypertension remain a mainstay of the prevention of immediate maternal complications and permanent cardiovascular injury, together with seizure prevention with magnesium sulfate.

White coat hypertension is reported in 25% of the nonpregnant adult population. Its prevalence in pregnancy is less certain, ranging from 4% to 30%.² According to 24-hour BP measurements, 32% of women with hypertension had white coat hypertension, but just 8% were diagnosed as such.¹¹¹ A meta-analysis of studies addressing white coat hypertension reported increased risks of preeclampsia and adverse fetal outcomes compared with women with normotension. Risks were lower compared with women with sustained chronic or gestational hypertension.⁸¹ The frequency and clinical significance of masked hypertension in pregnancy have not been extensively studied. Any category of nonsustained BP elevation in pregnancy can progress to sustained hypertension and requires follow-up. Self-measured BP is important for diagnosing nonsustained BP elevations, including masked hypertension and white coat hypertension, that occur before 20 weeks of gestation. For clinical purposes, the definition of hypertension in pregnancy requires 2 elevated BP measurements 4 hours apart (Table S1).

In the nonpregnant population, the association between BP variation, independent of baseline BP, and CVD risk is mixed, although greater variability is more convincingly associated with increased stroke risk.^112–118 Limited small studies of gestational short-term and visit-to-visit BP variation suggest that greater variation is associated with adverse maternal and perinatal outcomes,^119,120 but evidence is currently inconclusive, and there is need for consensus on the methodology for the measurement of BP variability in pregnancy.

Standard gestational age–specific BPs and centiles can assist in clinical interpretation of BP changes from expected levels.^53,120 Nationally representative, population-specific gestational BP references have been reported from China and the United Kingdom.^49,51 Studies addressing the association of BP changes in relation to healthy BP standards with maternal and perinatal outcomes are needed.

Most (≈90%) women with chronic hypertension have primary hypertension. Secondary hypertension may occur in a small proportion of women and is associated with worse maternal and fetal outcomes. It should be considered if maternal age is <35 years, hypertension is severe or resistant, there is no family history of hypertension, or there are suggestive laboratory features such as hypokalemia, elevated creatinine, or albuminuria early in pregnancy (Table S3).^121,122 Last, the prevalence of obesity in women of reproductive age has increased in recent years, and obstructive sleep apnea may play an increasing role in secondary hypertension among pregnant women.^123,124 Because there are no pregnancy-specific guidelines for obstructive sleep apnea treatment, pregnant women with sleep apnea should be managed concurrently with a sleep medicine specialist for application of available diagnostic and therapeutic methods, depending on the stage of pregnancy.

Postpartum hypertension and postpartum preeclampsia are not specifically included in the classification of HDP, but there is increasing awareness of their significance, as documented in the 2013 ACOG executive summary that implemented changes in clinical practice through closer postpartum monitoring and visits.¹²⁵ These entities are particularly important for 2 reasons. First, ≈60% of all maternal deaths occur within the first year postpartum, and HDP remain one of the leading causes of maternal mortality.¹²⁶ Second, postpartum hypertension offers an opportunity to use medications and to achieve BP goals without limitations related to their potential negative impacts on the fetus. The prevalence of postpartum hypertension may be as high as 8% in women without antepartum hypertension (followed up 48 hours after delivery and up to 6 weeks postpartum) and up to 50% in women with a history of preeclampsia 6 to 12 weeks after delivery.^127,128 The distinction between postpartum aggravation of antepartum HDP and de novo postpartum preeclampsia (also called delayed-onset postpartum preeclampsia) is unclear. Further research addressing underlying mechanisms is needed to clarify appropriate treatment and the need for magnesium sulfate for seizure prevention. The duration ranges from days to 3 months, contributing to serious short-term maternal complications such as stroke, seizures, and cardiomyopathy and metabolic dysregulation such as insulin resistance and weight gain.^127,128 Patient education is an important tool for early recognition of symptoms and signs. Novel approaches such as remote hypertension monitoring programs have potential to improve compliance and early diagnosis of postpartum hypertension and preeclampsia.¹²⁹

The rate of elevation in the antepartum soluble fms-like tyrosine kinase 1/placental growth factor ratio is an independent predictor of hypertension that persists postpartum.¹³⁰ Furthermore, preeclampsia-associated endothelial dysfunction and altered cerebrovascular autoregulation have been shown to persist postpartum¹³¹ and may amplify postpartum hypertension risk. Intravenous fluids, mobilization of extravascular fluid, and use of nonsteroidal anti-inflammatory drugs for postpartum analgesia may contribute to its occurrence. A recent randomized controlled clinical trial has shown that postpartum use of furosemide in women with HDP was associated with a 60% reduction in persistent hypertension at day 7 after delivery (adjusted relative risk, 0.40).¹³² If these findings can be implemented in the clinic, there is significant opportunity to reduce maternal morbidity in the postpartum period and to avoid unnecessary hospitalization. In nonpregnant individuals, there is abundant evidence that nonsteroidal anti-inflammatory drugs are associated with clinically significant increases in BP.^133–135 A recent systematic review and meta-analysis that included 5 randomized controlled trials and 5 retrospective cohorts concluded that compared with acetaminophen, nonsteroidal anti-inflammatory drugs were not associated with increased BPs up to discharge (2–4 days postpartum).¹³⁶ The authors considered the quality of evidence to be low because of the small sample sizes, imprecise results, and short duration of follow-up. Additional investigation is needed to address the impact of longer duration of postpartum nonsteroidal anti-inflammatory drug use in older women with chronic hypertension and additional renal and cardiovascular risk factors.^137,138

The recent American College of Cardiology/AHA task force guidelines lowered the threshold for the diagnosis of hypertension in nonpregnant patients to 130/80 mm Hg for stage 1 hypertension and to 140/90 mm Hg for stage 2 hypertension, resulting in larger numbers of individuals being diagnosed and treated.⁶ There is robust evidence in the general population demonstrating reduced CVD risk with treatment to lower levels.⁷ Indeed, most cardiovascular events occur in individuals with BP levels of 140 to 159/90 to 109 mm Hg.¹³⁹ Even younger individuals with hypertension demonstrate early vascular remodeling and endothelial dysfunction, particularly in smaller arteries and arterioles, which leads to progressive stiffening of larger blood vessels and organ damage if hypertension is untreated.^140,141 For all HDP, hypertension is defined internationally as a BP ≥140/90 mm Hg, although treatment thresholds and targets vary (Table 4).

The recommendations of published guidelines addressing diagnosis and treatment of HDP are summarized in Table 4. Differences among societies further demonstrate confusion in the field, which likely contributes to a failure to move forward. The ACOG recommends antihypertensive therapy for women with preeclampsia and a sustained systolic BP ≥160 mm Hg or diastolic BP ≥110 mm Hg and with chronic hypertension at a systolic BP ≥160 mm Hg or diastolic BP ≥110 mm Hg, with a treatment goal of 120 to 160/80 to 110 mm Hg.² Internationally, the majority of hypertension societies endorse a more aggressive approach for antihypertensive treatment, recommending therapy when BP is ≥140/90 mm Hg.^{110,142–144,149,151} Therapeutic targets similar to the American College of Cardiology/AHA target of 130/80 mm Hg⁶ are recommended by the International Society for the Study of Hypertension in Pregnancy,¹¹⁰ Hypertension Canada Guidelines,^146,149 National Institute for Health and Care Excellence,¹⁴⁴ and World Health Organization.¹⁴³ The following question arises: Why are the diagnostic and treatment BP thresholds higher in the United States compared with those recommended for nonpregnant individuals and compared with the majority of international guidelines addressing HDP?

Determining the optimal BP threshold in pregnancy for antihypertensive treatment and therapeutic targets requires a balance between the prevention of maternal hypertensive complications and the avoidance of fetal risks. The US (ACOG) guidelines are influenced by at least 3 debated issues. First is the prevailing perspective, based on small studies, that there are no measurable immediate or long-term health benefits of stricter BP treatment for the relatively short duration of pregnancy (4–9 months, depending on type of HDP) in young women without other CVD risks. Second, there are concerns that lowering maternal BP may compromise utero-placental circulation and negatively affect fetal well-being and growth. Third, therapeutic options are limited because of concerns about potential adverse fetal effects, particularly malformations from intrauterine exposure to antihypertensive medications. Furthermore, discrepancies among international guidelines are a reflection of the country-specific context within which they were developed. Such debate and subsequent inconsistencies in recommendations hinder progression toward consensus for optimal management of HDP internationally. For example, differences in BP thresholds for initiating antihypertensive therapy make combining results from observational studies of antihypertensive therapy for meta-analysis more challenging.

There are several compelling reasons to consider lower BP thresholds. First, more aggressive treatment of hypertension in pregnancy prevents the development of severe hypertension, as demonstrated by both a systematic review of randomized trials¹⁵² and CHIPS (Control of Hypertension in Pregnancy Study), in which the average BP achieved by tight control was 133/85 mm Hg.¹⁴ Although comparison of less tight versus tight control showed no effect on rates of preeclampsia, the former group demonstrated a higher risk of thrombocytopenia and elevated liver enzyme levels, markers of disease severity. In addition, in this trial and elsewhere, tight control may have decreased the risk of preterm birth.¹⁵³ The importance of severe hypertension as an outcome has been questioned,¹⁹ although exploratory analyses of the CHIPS data (adjusted for allocated group and prognostic factors) showed that severe hypertension is a surrogate marker for adverse maternal and perinatal outcomes, independently of and similar in magnitude to preeclampsia.^14,19 This is especially relevant in high-risk populations such as Black women for whom the risk of hypertension-related adverse outcomes is high.¹⁵⁴ A study of Black women with chronic hypertension showed that using antihypertensives before 20 weeks of gestation and achieving a BP <140/90 mm Hg were associated with lower incidences of superimposed preeclampsia and preterm delivery <35 weeks compared with women with BP ≥140/90 mm Hg.¹⁵⁵ Furthermore, lower (<140/90 mm Hg) versus higher (≥140/90 mm Hg) BP levels during pregnancy have been associated with lower rates of preeclampsia, including preeclampsia with severe features, and lower rates of preterm delivery.¹⁵³ Lower rates of preeclampsia with treatment of hypertension reported by these most recent studies are in sharp contrast to the majority of previous studies indicating that treatment of hypertension does not prevent preeclampsia. Whether there is a difference between women with chronic hypertension (who were preferentially recruited in these 2 studies indicating benefit) and those with gestational hypertension remains unknown; the answer will require prospective, adequately powered studies. On the basis of results of retrospective studies, including one showing benefit of tighter BP control¹⁵⁶ and another indicating that malignant/uncontrolled hypertension in the nonpregnant state has changes in the brain similar to those from eclampsia,¹⁵⁷ a large randomized controlled trial, the CHAP Project (Chronic Hypertension and Pregnancy), is nearing completion in the United States (ClinicalTrials.gov identifier: NCT02299414). The CHAP project is comparing outcomes between pregnant women with chronic hypertension who are given antihypertensive treatment to maintain BP <140/90 mm Hg with women given no treatment unless BP is ≥160/105 mm Hg.

Second, there is evidence that the pathophysiology of the neurological manifestations (headaches, visual disturbances, seizures) of preeclampsia is similar to that of the posterior reversible leukoencephalopathy syndrome.¹⁵⁸ Women with preeclampsia may be more susceptible to severe neurological outcomes such as intracerebral hemorrhage at lower systolic BPs (eg, 150–170 mm Hg)¹⁵⁷ compared with nonpregnant subjects, thus raising the possibility that lowering BP below current targets (eg, <150/90 mm Hg) may prevent these rare but devastating outcomes.¹⁵⁷

Third, treatment of nonsevere hypertension in pregnancy (eg, BPs 140–155/90–109 mm Hg) may permit prolongation of pregnancy in women without other severe features of preeclampsia who would require delivery.

Fourth, ACOG guidelines recommend withholding antihypertensive therapy for patients with preeclampsia unless BP approaches 160/110 mm Hg. They also recommend urgent delivery for women with severe features of preeclampsia, which include uncontrollable hypertension with BP ≥160/110 mm Hg, even for pregnancies <34 gestational weeks, unless high-level care is available in facilities with adequate maternal and neonatal intensive care resources.^3,110,159 Lowering thresholds for treatment may allow timely BP control and avoidance of rushed deliveries that commonly lead to prematurity and related complications.

Fifth, the classic view that young women with hypertension without other CVD risk factors are at low short-term CVD risk from untreated hypertension during the duration of pregnancy is challenged by current epidemiological and demographic trends toward advanced age at first pregnancy and higher CVD risk (subclinical or diagnosed).^9,160–162 This could also be relevant among women with multiple pregnancies, who may spend several years of their lives either pregnant or breastfeeding with uncontrolled hypertension. In addition, modern fertility techniques facilitate pregnancy in women with preexisting conditions associated with elevated CVD risk (eg, diabetes, chronic kidney disease, and polycystic ovary syndrome). Preexisting chronic kidney disease and heart disease are present in 3% and 1% to 4% of pregnancies in high-income countries, respectively.¹⁶³ Several guidelines consequently endorse more aggressive treatment in these women.^147,148

Finally, there is abundant evidence that HDP are associated with increased risk of both immediate and postpartum complications (such as acute cardiovascular and cerebrovascular disease)¹⁶⁴ and future maternal vascular disease (Table 2). Whether better management of BP during pregnancy will lead to lower rates of morbidity related to hypertension in the immediate postpartum period is not known. Traditional CVD risk factors (eg, obesity, hypertension, diabetes, hyperlipidemia) are associated with increased risk of HDP,⁷³ but the associations between HDP and future CVD, renal disease, and vascular dementia persist, even after adjustment for such factors.^31,34,36 It is estimated that approximately two-thirds of HDP-associated CVD risk is mediated by established risk factors, and the remainder is likely explained by an HDP-specific pathogenesis.^21,29 Whether treatment of nonsevere hypertension is beneficial for preventing long-term morbidity beyond pregnancy and the puerperium remains to be demonstrated. Furthermore, evidence is needed to clarify concerns about the observed, albeit nonstatistically significant, trend toward increased small-for-gestational-age risk and decreased preterm birth in women with tight versus less tight BP control in CHIPS.¹⁶⁵ The possible risk of drug-associated fetal malformations, long-term neurodevelopmental effects on offspring,¹⁶⁶ and suggested differential effects on these outcomes by antihypertensive class^152,167 are all areas that require further investigation.

Given new developments in the field of hypertension outside of pregnancy that support lower BP treatment targets, together with emerging data from larger clinical trials in pregnancy, this working group supports continued investigation to determine whether BP levels similar to those recommended outside of pregnancy for the initiation of therapy and as therapeutic targets are beneficial for the mother and safe and beneficial for the fetus. While awaiting more conclusive data and trials nearing completion, we endorse informed decision-making in partnership with the patient as to whether to treat nonsevere hypertension during pregnancy to targets similar to those recommended in nonpregnant individuals. Personalization of therapy, by giving special attention to other risk factors related to hypertension-related adverse outcomes (such as preexisting heart or kidney disease, obesity, and Black race), is a rational approach.

Initial antihypertensive therapy is widely established to be monotherapy with an accepted first-line drug: labetalol or methyldopa. Some,^{1,110,142–144,147,149} but not all,¹⁴⁸ societies support the use of nifedipine as an initial therapy. In countries where labetalol is unavailable (eg, Germany), alternative β-blockers such as metoprolol or oxprenolol can be considered. These therapeutic options are based on small individual trials and are advocated by national and international clinical practice guidelines. There is no clear evidence that one drug is preferable to another according to a systematic review of randomized trials for all types of pregnancy hypertension considered together, for all antihypertensives considered together, or for β-blockers (including labetalol) considered separately.¹⁵² However, in a separate network meta-analysis, specifically for treatment of chronic hypertension, atenolol was associated with fetal growth restriction,¹⁶⁸ especially when given for a longer duration.¹⁶⁹ These data conflict with some observational studies that have associated β-blocker treatment (including labetalol) with an excess of small-for-gestational-age infants, although the authors did not necessarily adjust for treatment indication and severity of maternal disease.¹⁷⁰ These conflicting data underscore a need for more fetal and newborn data on the safety of currently used antihypertensive agents in pregnancy.

Numerous clinical trials have compared various short-acting antihypertensives in the setting of acute, severe hypertension in pregnancy. The drugs most commonly examined are parenteral hydralazine, parenteral labetalol, and oral nifedipine (short, intermediate, or long acting). A Cochrane review concluded that these drugs were comparable with respect to safety and efficacy and recommended that professionals choose on the basis of experience and familiarity with a particular drug.¹⁷¹ Most cases of severe hypertension can be successfully controlled with these drugs using doses and protocols recommended by professional societies.¹⁷² In resource-poor countries, a report documented successful treatment of acute severe hypertension with oral preparations of labetalol, intermediate-acting nifedipine, and methyldopa.¹⁷³ Additional agents that may be considered for resistant hypertension, although not extensively studied, include nicardipine, clonidine, and furosemide.^174–176 Notably, diuretics, the mainstay of hypertension treatment in nonpregnant individuals, are not used often in pregnant women. This position is informed mainly by earlier studies suggesting that women with preeclampsia have lower plasma volume, suggesting that diuretics may further aggravate volume depletion and promote reactive vasoconstriction. However, older studies demonstrated their favorable safety profile in pregnancy,¹⁷⁷ and more recent guidelines have acknowledged that in women with salt-sensitive chronic hypertension or chronic kidney disease and reduced glomerular filtration rate, diuretics may be used safely, although perhaps at lower doses.² Recent studies demonstrate that they may be particularly effective in postpartum hypertension.¹³²

The limitations of existing data on the safety of antihypertensives in pregnancy are highlighted by a systematic review of studies addressing in utero exposure to antihypertensive medications and adverse fetal outcomes. Only 5 of 47 studies were considered high quality, and few studies reported increased odds of adverse effects in treated compared with normotensive untreated women, including congenital malformations, and effects were not uniformly observed across different studies using the same medications.¹⁶⁶ Furthermore, similar adverse events have been reported in untreated women with hypertension, leading to the conclusion that the evidence for teratogenicity of most antihypertensive agents is weak.^178,179

Although first-trimester exposure to medication raises concern about structural malformations (other than those attributable to physical or vascular disruption), the fetal central nervous system develops throughout gestation and may be affected by exposures at any time point. However, no firm conclusions can be drawn about long-term child outcomes given the paucity of relevant high-quality studies.¹⁶⁶ No adverse neurodevelopmental effects have been observed for methyldopa,¹⁸⁰ nifedipine,¹⁸¹ or atenolol,¹⁸² although atenolol should be used with caution (see above). Registry data adjusted for important covariates were reassuring about the effects of preeclampsia itself; only a minimal effect was seen on standardized mathematics test scores in children from affected pregnancies at 9, 12, and 15 years of age.¹⁸³ In addition, when control subjects with untreated or treated hypertension have been used, children from labetalol- and methyldopa-treated women had similar IQ scores.^{2,14,158,184,185} Small clinical trials and observational studies suggest that amlodipine, clonidine, and thiazide diuretics are probably safe in pregnancy as well.^177,186,187 It is also widely accepted that all renin-angiotensin system blockers should be avoided during pregnancy,¹⁸⁸ especially during the second and third trimesters when blockade of the fetal renin-angiotensin system clearly interferes with kidney development and function. Given suboptimal and frequently contradicting data on fetal safety after exposure to antihypertensive medications in utero, well-designed, carefully controlled trials are needed, with attention given to short- and long-term fetal and maternal outcomes. Last, the professionals in the field should be familiar with services offered by the Organization of Teratology Information Specialists.¹⁸⁹ The organization was founded in 1987 as a way of connecting experts in the field of birth defects research to the general public. It provides up-to-date information about the risks of medications during pregnancy and breastfeeding to patients, health care professionals, and researchers in the field of teratology.