Blood Volume

Very early in the first trimester, pregnant women experience an expansion of renal blood flow and glomerular filtration rate (GFR). Filtered sodium increases by about 50%. Despite physiologic changes that would promote loss of salt and water and contraction of blood volume, the pregnant woman will expand her blood volume by 40% to 50%. In part, the stimulation to retain fluid may be a response to the fall in vascular resistance and reduction in blood pressure. The renin-angiotensin system is activated, and the plasma concentration of aldosterone is elevated. Although the simplicity of this explanation is attractive, the actual process is probably much more complicated.

As plasma volume expands, the hematocrit falls, and hematopoiesis is stimulated. Red cell mass will expand from 18% to 25% depending on the status of individual iron stores. Physiologic anemia with a maternal hematocrit between 30% and 35% does not usually complicate pregnancy in the context of maternal heart disease. More significant anemia, however, may increase cardiac work and induce tachycardia. Microcytosis due to iron deficiency may impair perfusion of the microcirculation of patients who are polycythemic because of cyanotic heart disease; this is because microcytic red blood cells are less deformable. Iron and folate supplementation may be appropriate.

In a similar fashion, serum albumin concentration falls by 22% despite an expansion of intravascular albumin mass by 20%. As a result, serum oncotic pressure falls in parallel by 20% to about 19 mm Hg. In normal pregnancy, intravascular fluid balance is maintained by a fall in interstitial oncotic pressure. However, if LV filling pressure becomes elevated or if pulmonary vascular integrity is disrupted, pulmonary edema will develop earlier in the disease process than in nonpregnant women.

Risk-Scoring Strategies

Pregnancy in women with heart disease is associated with increased risks for deterioration of maternal cardiac status and adverse pregnancy outcomes. These risks include maternal arrhythmias, heart failure, preterm birth, fetal growth restriction, and a small but significant risk of maternal and fetal mortality. Accurate quantification of maternal and fetal risks should be used to counsel patients and to direct care. Three risk models have been suggested.

The Cardiac Disease in Pregnancy (CARPREG) score was derived from a prospective descriptive study of 562 pregnant women with cardiac disease that included congenital or acquired lesions and arrhythmias. The scoring system was created to estimate the risk of experiencing a primary cardiac event. The predictors in this scoring system include (1) a prior cardiac event—heart failure, transient ischemic attack, or stroke before pregnancy; (2) baseline New York Heart Association (NYHA) class greater than II or cyanosis; (3) mitral valve area less than 2 cm ² , aortic valve area less than 1.5 cm ³ , or peak LVOT gradient greater than 30 mm Hg by echocardiography; and (4) reduced systemic ventricular systolic function with an ejection fraction (EF) of less than 40%.

The ZAHARA ( Zwangerschap bij Aangeboren Hartafwijkingen [Pregnancy In Congenital Heart Disease]) score was derived from a nationwide database of 1302 pregnant women with CHD. Predictors identified to be associated with maternal cardiac complications included (1) prior arrhythmia, (2) NYHA class III or IV, (3) LVOT gradient greater than 50 mm Hg or aortic valve area less than 1.0 cm ² , (4) mechanical valve prosthesis, (5) systemic atrioventricular (AV) valve regurgitation (moderate/severe), (6) pulmonary AV valve regurgitation (moderate to severe), (7) cardiac medication prior to pregnancy, and (8) cyanotic heart disease, both corrected and uncorrected. This study also externally validated the prior CARPREG study and noted that the CARPREG score overestimated the risk.

The modified World Health Organization (WHO) classification uses four categories determined largely by diagnosis: class I includes uncomplicated, mild pulmonary stenosis; class II comprises unoperated ASD, ventricular septal defect (VSD), and repaired tetralogy of Fallot; class III includes mechanical valves, systemic right ventricle, Fontan circulation, unrepaired cyanotic heart disease, other complex CHD, Marfan syndrome with an aorta 40 to 45 mm in width or bicuspid aortic valve with an aorta 45 to 50 mm; and class IV describes pulmonary hypertension/Eisenmenger syndrome, systemic EF less than 30%, systemic dysfunction of NYHA class III or IV, severe mitral stenosis, severe symptomatic aortic stenosis, Marfan syndrome with aorta greater than 45 mm, bicuspid valve with aorta greater than 50 mm, or severe coarctation.

The ZAHARA II study was performed to validate and compare CARPREG, ZAHARA I, and the modified WHO risk models for pregnant women with CHD. ZAHARA II enrolled 213 women and included patients only with congenital structural heart disease. Overall, primary cardiovascular events were observed in 22 of the pregnancies (10.3%). The most frequent events included clinically significant arrhythmias, followed by heart failure and thrombotic events. It was noted that the ZAHARA I and CARPREG scores overestimated risk. The modified WHO classification performed as the best available risk-assessment model for estimating cardiovascular risk.

From the scoring systems, common features can be identified that individually or collectively may predict adverse outcome. For patients with several risk factors, the impact on outcome may be more than additive, as suggested by the structures of the scoring systems: (1) prior cardiac event, (2) NYHA class III or IV, (3) LVOT obstruction, (4) reduced systemic EF, (5) mechanical prosthesis, (6) moderate to severe AV valve regurgitation, (7) cardiac medications prior to pregnancy, and (8) cyanotic heart disease. The WHO system based on diagnoses incorporates risk associated with pulmonary hypertension and RV dysfunction, severe mitral stenosis, dilated aortas, and single-ventricle repairs not specifically captured by CARPREG or ZAHARA.

Although it seems optimal to be able to assign a direct risk score to a woman when counseling her during pregnancy, all of the studies above highlight the challenge in being able to create the ideal scoring system. Each category of CHD carries varied risks based on maternal hemodynamic and cardiovascular function findings. It is therefore important to understand the different risk scores; although ultimately, the individual cardiovascular functional parameters and overall hemodynamic stability of each patient must be understood to provide the most individualized care for each woman.

Mitral Stenosis

Mitral stenosis is most commonly caused by rheumatic heart disease and is the most common acquired valvular lesion in pregnant women. Valvular dysfunction progresses continuously throughout life. Deterioration may be accelerated by recurrent episodes of rheumatic fever, an immunologic response to group A β-hemolytic Streptococcus (GBS) infections. The incidence of rheumatic fever in a population is heavily influenced by conditions of poverty and crowding. These same individuals are at risk for having reduced access and use of health care resources and may present undiagnosed or untreated.

Patients with asymptomatic mitral stenosis have a 10-year survival rate of greater than 80%. Once a patient is significantly symptomatic, the 10-year survival rate without treatment is less than 15%. In the presence of pulmonary hypertension, mean survival falls to less than 3 years. Death is due to progressive pulmonary edema, right-sided heart failure, systemic embolization, or pulmonary embolism.

Stenosis of the mitral valve impedes the flow of blood from the left atrium to the left ventricle during diastole. The normal mitral valve area is 4 to 5 cm ² . Symptoms with exercise can be expected with valve areas less than or equal to 2.5 cm ² . Symptoms at rest are expected at less than or equal to 1.5 cm ² . The left ventricle responds with Starling mechanisms to increased venous return with increased performance, elevating CO in response to demand. The left atrium is limited in its capacity to respond, and therefore CO is limited by the relatively passive flow of blood through the valve during diastole; increased venous return results in pulmonary congestion rather than increased CO. Thus the drive for increased CO in pregnancy cannot be achieved, resulting in increased pulmonary congestion. The relative tachycardia experienced in pregnancy shortens diastole, decreases LV filling, and therefore further compromises CO and increases pulmonary congestion.

The diagnosis of mitral stenosis in pregnancy before maternal decompensation is uncommon. Fatigue and dyspnea on exertion are characteristic symptoms of mitral stenosis but are also ubiquitous among pregnant women. Although the presence of a diastolic rumble may suggest mitral stenosis, this finding is subtle and may be overlooked or not appreciated. Not uncommonly, an intercurrent event such as a febrile episode will result in exaggerated symptoms and the diagnosis of pulmonary edema or oxygen desaturation. Under these circumstances, particularly in the context of a patient from an at-risk group, an echocardiogram should be performed to rule out mitral valvular disease.

Echocardiographic diagnosis of mitral stenosis is based on the characteristic appearance of the stenotic, frequently calcified valve. Calculation of valve area from the Doppler pressure half-time method or by 2-D planimetry provides an objective measure of severity. Valve areas of 1 cm ² or less usually require pharmacologic management during pregnancy and invasive hemodynamic monitoring during labor, whereas those 1.4 cm ² or less usually require careful expectant management. Left atrial enlargement identifies a patient at risk for atrial fibrillation, subsequent atrial thrombus, and the potential for systemic embolization. Embolic complications have been reported in pregnant women with atrial enlargement without atrial fibrillation. Pulmonary hypertension, a complication of worsening mitral disease, can be diagnosed and quantified with Doppler evaluation of the regurgitant jet across the tricuspid valve. Elevated pulmonary pressures may be due to hydrostatic forces associated with elevated left atrial pressures or, in more advanced disease, may result from pathologic elevations of pulmonary vascular resistance (PVR). Hydrostatic pulmonary hypertension may respond to therapy that lowers left atrial pressure. Pulmonary hypertension due to elevated PVR is life threatening in pregnancy and may precipitate right-sided heart failure in the postpartum period.

Pregnancy does not negatively affect the natural history of mitral stenosis. Chesley reviewed the medical histories of 134 women with functionally severe mitral stenosis who survived pregnancies between 1931 and 1943. These women lived before modern management of mitral stenosis and therefore represent the natural history of the disease. By 1974, only nine of the cohort remained alive. Their death rate was exponential; during each year of follow-up, the rate for the remaining cohort was 6.3%. Women with subsequent pregnancies had comparable survival to those who did not become pregnant again, allowing the authors to conclude that pregnancy did not negatively affect long-term outcome.

The goal of antepartum care in the context of mitral stenosis is to achieve a balance between the drive to increase CO and the limitations of flow across the stenotic valve. Most women with significant disease require diuresis with a drug such as furosemide. In addition, β-blockade reduces heart rate, improves diastolic flow across the valve, and relieves pulmonary congestion. Al Kasab and associates evaluated the impact of β-blockade on 25 pregnant women with significant mitral stenosis. Figure 37-5 describes the functional status of women before pregnancy and during pregnancy before and after β-blockade. The deterioration associated with pregnancy and the subsequent improvement with treatment is evident. Fastidious antepartum care as described earlier should supplement pharmacologic management.

The effects of β-blockade on functional status of women with mitral stenosis. NYHA, New York Heart Association.

(From Al Kasab S, Sabag T, Al Zaibag M, et al. Beta-adrenergic receptor blockade in the management of pregnant women with mitral stenosis. Am J Obstet Gynecol. 1990;163:37.)

Women with a history of rheumatic valvular disease who are at risk for contact with populations with a high prevalence of streptococcal infection should receive prophylaxis with daily oral penicillin G or monthly benzathine penicillin. Most pregnant women live in close contact with groups of children and usually are considered at risk.

Atrial fibrillation (AF) is a complication associated with mitral stenosis due to left atrial enlargement. Rapid ventricular response to AF may result in sudden decompensation. Digoxin, β-blockers, or calcium channel blockers can be used to control ventricular response. In the context of hemodynamic decompensation, electrical cardioversion may be necessary. Anticoagulation with heparin should be used before and after cardioversion to prevent systemic embolization. Patients with chronic AF and a history of an embolic event should also undergo anticoagulation. Anticoagulation may be considered in women with a left atrial dimension of 55 mm or greater.

Labor and delivery can frequently precipitate decompensation in patients with critical mitral stenosis. Pain induces tachycardia, and uterine contractions increase venous return and thereby increase pulmonary congestion. Women with critical mitral stenosis frequently cannot tolerate the work of pushing in the second stage. Clark and coworkers described the abrupt elevation in pulmonary artery pressures in the immediate postpartum period associated with return of uterine blood to the general circulation ( Fig. 37-6 ). Aggressive, anticipatory diuresis will reduce pulmonary congestion and the potential for oxygen desaturation.

The changes in pulmonary capillary wedge pressure (PCWP) associated with delivery and subsequent diuresis in women with mitral stenosis. PP, postpartum.

(From Clark S, Phelan J, Greenspoon J, et al. Labor and delivery in the presence of mitral stenosis: central hemodynamic observations. Am J Obstet Gynecol. 1985;152:384.)

The hemodynamics of women with symptomatic stenosis or a valve area of 1 cm ² or less may benefit from management with the aid of a pulmonary artery catheter. Ideally, hemodynamic parameters are assessed when the patient is well compensated, early in labor. These findings serve as a reference point to guide subsequent therapy. Pain control is best achieved with an epidural, and heart rate control is maintained through pain control and β-blockade. To avoid pushing, the second stage is shortened with low forceps or vacuum delivery; cesarean delivery is reserved for obstetric indications. Aggressive diuresis is initiated immediately postpartum. In a series of 80 pregnancies managed with a range of severity, the most common complications were pulmonary edema (31%) and arrhythmia (11%). When valve area was 1 cm ² or less, the rate of pulmonary edema was higher (56%), as was the rate of arrhythmia (33%). These rates will be dependent on the effectiveness of medical management and the timing of presentation and diagnosis.

Aggressive medical management, including hospital bed rest in selected cases, is sufficient to manage most women with mitral stenosis. The woman with uncommonly severe disease may require surgical intervention. Although successful valve replacement and open commissurotomy have been reported in pregnancy, they are now rarely needed. Two reports detail successful balloon valvotomy in a series of 40 and 71 women with minimal complications. Complications of balloon valvuloplasty outside of pregnancy occur at the following rates: mortality (0.5%), cerebrovascular accident (1%), and mitral regurgitation that requires surgery (2%). Mitral valvuloplasty in pregnancy has been reported with success rates of 95% or greater. The incidence of severe mitral regurgitation that required surgery was 4.6% in the larger series at long-term follow-up; no women required surgery for acute mitral regurgitation during pregnancy. The rate of fetal loss was between 1% and 2%. Medical management should be clearly exhausted before assuming these risks during pregnancy, when emergent intervention such as valve replacement is more complicated and carries a significant risk to the fetus.

Rheumatic disease can also affect the aortic valve. In the context of aortic stenosis that is critically dependent on ventricular filling, management of significant mitral stenosis that limits ventricular filling is particularly complicated.

Mitral Regurgitation

Mitral regurgitation may be due to a chronic progressive process such as rheumatic valve disease or myxomatous mitral valve disease, frequently associated with mitral valve prolapse. As regurgitation increases over time, forward flow is maintained at the expense of LV dilation with eventual impaired contractility. Left atrial enlargement may be associated with AF that should be managed with ventricular rate control and anticoagulation. The patient with chronic mitral regurgitation may remain asymptomatic even with exercise. Preconceptional counseling should include consideration of valve replacement in consultation with a cardiologist. In general, valve replacement is recommended for (1) symptomatic patients, (2) AF, (3) EF less than 60%, (4) LV end-diastolic dimension greater than 40 mm, or (5) pulmonary systolic pressure greater than 50 mm Hg. As discussed later, the benefits of valve replacement before pregnancy must be balanced against the risks associated with a prosthetic valve in pregnancy and the potential for prosthetic valve deterioration in pregnancy. If surgery is required, valve repair—rather than replacement—is preferred when possible to avoid the need for anticoagulation.

Acute mitral regurgitation in young patients is uncommon and may be associated with ruptured chordae tendineae as a result of endocarditis or myxomatous valve disease. Without time for LV compensation, forward flow may be severely compromised; urgent valve surgery is usually required. Inotropic LV support and systemic afterload reduction can be used to stabilize the patient.

The hemodynamic changes associated with pregnancy can be expected to have mixed effects. A reduction in systemic vascular resistance (SVR) tends to promote forward flow. The drive to increase CO will exacerbate LV volume overload, and increased atrial dilation may initiate AF. Pulmonary congestion can be managed by careful diuresis with the knowledge that adequate forward flow is usually dependent on a high preload to achieve adequate LV filling. AF should be managed as in the nonpregnant state. An increase in SVR due to progressive hypertension secondary to advancing preeclampsia may significantly impair forward flow and should be treated. Labor and delivery should be managed with standard cardiac care. Catecholamine release that occurs as a result of pain or stress impairs forward flow; therefore particular attention should be paid to LV filling. Excessive preload results in pulmonary congestion, and insufficient preload will not fill the enlarged left ventricle and will result in insufficient forward flow. A pulmonary artery catheter can be used to determine appropriate filling pressure in early labor or before induction. Although a large v-wave may complicate the interpretation of pulmonary artery wedge pressure, the pulmonary artery diastolic pressure can be used as a reference point. Diuresis in the early postpartum period may be required.

Myxomatous mitral valve disease or mitral valve prolapse is a common condition that affects as many as 12% of young women. In the absence of conditions of abnormal connective tissue such as Marfan or Ehlers-Danlos syndrome and clinically significant mitral regurgitation, women with mitral prolapse can be expected to have uncomplicated pregnancies. They may experience an increase in tachyarrhythmias that can be treated with β-blockers, and the use of prophylactic antibiotics may be considered at the time of delivery.

Aortic Stenosis

Most patients who develop calcific stenotic trileaflet aortic valves do so outside their childbearing years (age 70 to 80 years). Patients with bicuspid valves develop significant stenosis after the age of 50 to 60 years. Rheumatic disease can also affect the aortic valve, usually after the development of significant mitral disease. Most pregnant women with significant aortic stenosis have congenitally stenotic valves: bicuspid valves with congenitally fused leaflets, unicuspid valves, or tricuspid valves with fused leaflets.

The natural history of aortic stenosis is characterized by a long, asymptomatic period. With increasing outflow obstruction, patients develop angina, syncope, and LV failure. Without valve replacement, only 50% of patients will survive 5 years after the development of angina; 3 years after the development of syncope; and 2 years after the development of LV failure. Although valve replacement is the only definitive treatment for calcific aortic stenosis, valvuloplasty may prove beneficial in some young adults whose valves are not calcified. Medical management of symptomatic patients is not generally efficacious. Mechanical valve replacement requires anticoagulation, which complicates subsequent pregnancies.

Young women with aortic stenosis are usually asymptomatic. Although they may develop increasing exercise intolerance in pregnancy, the progression is insidious and not easily distinguished from the effects of normal pregnancy. The diagnosis is usually made by the auscultation of a harsh systolic murmur. The murmur can easily be distinguished from a physiologic murmur of pregnancy by its harshness and radiation into the carotid arteries. Diagnosis is confirmed by echocardiography whereby the pressure gradient across the valve can be measured by Doppler, and the valve area can be calculated with the continuity equation. Many women with significant aortic stenosis experience the expected increase in CO associated with pregnancy. Increased flow across the fixed, stenotic valve results in a proportionately increased gradient across the valve. Although the pressure gradient during pregnancy may be higher than that observed postpartum, these differences are not significant.

Four series of patients with aortic stenosis in pregnancy have been reported. The reports summarize experiences with wide ranges in severity of disease and management ranging from the 1960s and 1970s to the present. Arias and Pineda described a series of 23 cases managed before 1978 with a maternal mortality rate of 17%. More recent series, however, do not demonstrate this high level of maternal risk. The potential for serious adverse outcomes reported by Arias should, however, serve as an indication for intensive management. The rate of mortality should not necessarily be used as an indication for termination or surgical intervention. Pregnant patients have been successfully managed with aortic gradients in excess of 160 mm Hg. In general, patients with a peak aortic gradient of 60 mm Hg or less have had uncomplicated courses. Those with higher gradients require increasingly intensive management.

Aortic valve replacement and balloon valvotomy have been reported during pregnancy. Balloon valvotomy in a young patient without valve calcification can provide significant long-term palliation. Valvotomy before pregnancy may provide an interval of hemodynamic stability sufficient to complete a pregnancy without the complications associated with a mechanical prosthetic valve. Consideration for valve replacement or valvotomy during pregnancy should be reserved for patients who remain clinically symptomatic despite hospital care. In general, intervention should not be based solely on a pressure gradient or valve area.

Aortic stenosis is a condition of excess LV afterload. Ventricular hypertrophy increases cardiac oxygen requirements, whereas increased diastolic ventricular pressure impairs coronary perfusion. Each increases the potential for myocardial ischemia. The left ventricle requires adequate filling to generate sufficient systolic pressure to produce flow across the stenotic valve. Given a hypertrophied ventricle and some degree of diastolic dysfunction, the volume-pressure relationship is very steep. A small loss of LV filling results in a proportionately large fall in LV pressure and, therefore, a large fall in forward flow and CO. The pregnant patient with significant aortic stenosis is very sensitive to loss of preload associated with hemorrhage or epidural-induced hypotension. The window of appropriate filling pressure is narrow. Excess fluid may result in pulmonary edema; insufficient fluid may result in hypotension and coronary ischemia. In general, pulmonary edema associated with excess preload is much easier to manage than hypotension due to hypovolemia.

Appropriate antepartum care is described earlier. Given that most aortic stenosis in young women is congenital in origin, fetal echocardiography is indicated. Although some controversy persists, cesarean delivery is generally reserved for obstetric indications. Pain during labor and delivery can be safely managed with regional analgesia using a low-dose bupivacaine and narcotic technique. Dense anesthesia during the second stage can be obtained with minimal hemodynamic complications using a caudal catheter. Patients with gradients above 60 to 80 mm Hg may benefit from the use of a pulmonary artery and arterial catheter during labor. Hospital admission one day before planned induction of labor with a favorable cervix is preferred, and a prolonged induction should be avoided. Pulmonary artery and radial artery catheters, as well as epidural and caudal catheters, are placed. The patient should be gently hydrated overnight to achieve a pulmonary artery wedge pressure (PAWP) of 12 to 15 mm Hg. Some patients with milder disease spontaneously diurese in the face of a volume load such that an elevation in PAWP cannot be achieved. An elevated PAWP serves as a buffer against a loss of preload. If PAWP falls with bleeding or the onset of anesthesia, volume can be administered before a reduction in forward flow occurs. In general, pushing is minimized, and the second stage is shortened with operative vaginal delivery. Antibiotics may be considered for the prevention of endocarditis.

Postpartum patients should be monitored hemodynamically for 24 to 48 hours. Diuresis is usually spontaneous, and the patient can be allowed to find her predelivery compensated state. When diuresis must be induced to treat pulmonary edema, it should be done gently and carefully. Predelivery hemodynamic parameters should be used as an end point. Some have found that a significant delay in valve replacement in women with quite severe disease is associated with maternal complications. In a larger cohort of women with less severe disease followed for 6 years and compared with a matched cohort who had not been pregnant, women who experienced a pregnancy had a reduction in event-free survival. These observations may be the result of accelerated valve deterioration due to pregnancy. For this reason, valve replacement within weeks of delivery may be indicated.

Aortic Regurgitation

Aortic regurgitation is most often due to a congenitally abnormal valve. Other causes include Marfan syndrome, endocarditis, and rheumatic disease. As with mitral regurgitation, the left ventricle compensates for decreased forward flow with an increase in LV end-diastolic volume. Afterload reduction prevents progressive LV dilation and is recommended for patients with LV dysfunction or dilation. Valve replacement is generally recommended for (1) NYHA functional class III and class IV symptoms, (2) an EF less than 50%, or (3) an LV end-systolic dimension greater than 50 mm. Acute regurgitation may be due to aortic root dissection or endocarditis and usually represents a medical emergency that requires urgent valve replacement.

The reduction in vascular resistance associated with pregnancy tends to improve cardiac performance. If afterload reduction has been achieved with an angiotensin-converting enzyme (ACE) inhibitor or an angiotensin receptor blocker (ARB) before pregnancy, hydralazine or a calcium channel blocker such as nifedipine should be substituted. Modest heart rate elevations should be tolerated, and bradycardia may be associated with increased regurgitation due to prolongation of diastole. Labor and delivery are managed with standard cardiac care, and pulmonary artery catheterization is not usually required. As the hemodynamic changes associated with pregnancy resolve, a rise in vascular resistance should be anticipated and afterload reduction maintained.

Prosthetic Valves

Definitive therapy for significant valvular disease requires surgical repair or, more commonly, replacement. Mechanical valves are durable but require anticoagulation. When used in a young woman, bioprosthetic valves usually require replacement during her lifetime. Reports of pregnancies associated with prosthetic valves suggest significant variability in outcomes, and these have been reviewed in detail by Elkayam and Bitar and within the 2014 AHA/ACC Valvular Heart Disease Guidelines. Reported outcomes must be interpreted in the context of the cohorts of patients reported and the circumstances of clinical care.

Decisions that surround the timing and choice of valve replacement for a woman of reproductive age are complex. Managing a pregnancy with moderate valve disease may be less complicated than managing a pregnancy with a prosthetic valve. The durability of a mechanical valve has considerable advantages for a young person, but it is associated with more adverse outcomes in pregnancy. Delay in valve replacement until childbearing is completed is appropriate when the severity of heart disease is believed to be manageable in pregnancy.

Bioprosthetic valves have relatively low rates of complications in pregnancy. Some women, particularly those with CHD, will have residual hemodynamic issues associated with their primary condition that are not addressed with valve replacement. The impact of pregnancy on the life of a bioprosthetic valve has been studied. Ten-year graft survival following two pregnancies was 16.7%, compared with 54.8% following a single pregnancy, which suggests that pregnancy may adversely affect the life of a bioprosthetic valve. Accelerated deterioration of bioprosthetic valves in the setting of pregnancy has been confirmed by several studies.

Anticoagulation is required with a mechanical valve. Man­agement of anticoagulation in pregnant women with mechanical prosthetic valves remains very controversial because commonly used anticoagulants have significant maternal and fetal adverse effects, and no single agent is safe throughout all stages of pregnancy. Interpretation of reported outcomes must be made in the context of valve location, thrombogenicity of the particular valve, strategies for anticoagulation and monitoring of effectiveness, and social context including compliance. Mechanical valves in the mitral position will be expected to have more thrombotic complications than those in the aortic position. Older-generation valves, such as Björk-Shiley or Starr-Edwards, may be in place in a pregnant woman and will be more likely to have thrombotic complications. When strategies for anticoagula t ion without dose adjustment are used in pregnancy, particularly in the case of heparin therapy, increased thrombotic complications are to be expected. Outcomes are better when clinical care teams are experienced with monitoring and dosing anticoagulation and work to achieve compliance.

Recommendations regarding anticoagulation in pregnancy for women with mechanical valves have been published by the ACC and AHA, the American College of Chest Physicians, and by Elkayam and Bitar. Each have largely drawn information from the same sources but in doing so have made substantially different recommendations regarding the use of warfarin or heparin. Physicians who care for pregnant women with mechanical heart valves should be familiar with each set of recommendations and should use them for guidance in counseling patients and establishing a plan.

Ideally, patients are evaluated and counseled preconceptionally. An effective plan of birth control is used until a pregnancy is planned; therefore long-acting reversible contraceptives should be considered. Progestin-based systems may reduce menstrual bleeding for women on warfarin. Once a pregnancy is desired, a clear plan for anticoagulation in the first trimester should be in place, with clinical systems available to immediately implement the plan once a pregnancy is identified. Early surveillance for a potential pregnancy is essential.