Introduction
Maternal safety surrounding nonobstetrical surgery, labor, and delivery has been seen as one of the greatest triumphs of 20th century medicine. Within living memory, nearly 1% of women in the United States died during or in the immediate aftermath of childbirth. As recently as 1900, maternal mortality was six to nine per 1000 live births, and of those 1000 births, 100 infants died before the age of 1 year. From 1900 through 1997, maternal mortality declined almost 99%, to less than 0.1 reported deaths per 1000 live births (7.7 deaths per 100,000 live births in 1997). Environmental interventions, improvements in nutrition, advances in clinical medicine, improvements in access to health-care, improvements in surveillance and monitoring of disease, increases in education levels, and improvements in standards of living contributed to this remarkable decline.
However, despite these significant improvements, a startling trend has arisen. Between 2000 and 2014, maternal mortality has steadily risen in the United States; the United States now has the highest maternal mortality among industrialized nations and is one of the only such nations to have a maternal mortality that continues to climb ( Fig. 29.1 ). Clearly, significant preventable causes of maternal death still persist. Here, we attempt to offer recommendations toward what authorities consider best practice in an effort to stem the tide and improve maternal morbidity and mortality in the perioperative and peripartum periods.
Causes of Maternal Morbidity and Mortality
Since 1952, the most thorough and reliable data on maternal mortality have come from the Confidential Enquiry into Maternal Death, now called Confidential Enquiry into Maternal and Child Health (CEMACH), and the Centers for Disease Control’s Maternal Mortality Survey. However, more recent data are available from the World Health Organization, with the most recent period covering data from 2003–2009 ( Fig. 29.2 ). The WHO reviewed a total of over 60,000 deaths from 115 countries. Among both developed and developing countries, obstetric hemorrhage remains the most common direct cause of maternal death, with rates of 16.3% and 27.1% of deaths respectively. The majority of hemorrhage deaths continue to occur in the postpartum period and do not appear to be decreasing in frequency worldwide. In developed nations, embolic events were the second-leading direct cause of death with hypertensive diseases third (13.8% and 12.9% respectively), and in developing regions, hypertensive diseases and sepsis (14% and 10.7%) were second and third. In both developed and d;eveloping regions, indirect causes of maternal death, primarily preexisting medical conditions, comprise a large proportion of all deaths (24.7% and 27.5% respectively).
Anesthetic Causes of Morbidity and Mortality
Serious anesthesia-related complications are thankfully rare in the perioperative and peripartum period. Historically, perioperative mortality fell significantly through the 1980s and 1990s, although it regressed slightly in the beginning of the 2000s to a rate of about 1 in 20,000. The overwhelming majority of these deaths were a result of failures in airway management: unidentified esophageal intubation and hypoventilation. Obesity, aspiration, and the lack of capnography were commonly associated with these deaths.
With the expansion of regional anesthesia use in these patients, the risk of anesthesia mortality has decreased significantly. The Society for Obstetric Anesthesia and Perinatology’s Serious Complications Repository Project reviewed more than 257,000 anesthetics between 2003 and 2009 and identified 157 serious complications, 85 of which were deemed to be anesthesia-related ( Fig. 29.3 ). No anesthesia-related deaths were reported in this group; however, comparatively common severe complications included failed intubation, with an incidence of 1:533 and high neuraxial block, with an incidence of 1:4336. Other routinely discussed complications of obstetric anesthesia were significantly less common: serious neurologic injury, with an incidence of 1:35,923, epidural abscess/meningitis, with an incidence of 1:62,866, and epidural hematoma, with an incidence of 1:251,423 are all comparatively rare.
Thromboembolism and Thrombophilia
Pregnancy and its hormonal changes are known to promote coagulation. Clotting factors increase with gestational age and, together with rapid myometrial contraction at delivery, help prevent excessive blood loss. Platelet production, consumption, and activation increase. Thromboelastography demonstrates accelerated clot formation, which further increases the risk of thromboembolism in pregnancy. Risk factors associated with maternal thromboembolism include older age, immobility, prolonged travel, surgery, family history, patient history, oral contraceptive use, and obesity. Risk factors for thromboembolism in pregnancy are outlined in Fig. 29.4 .
With the continually significant risk of embolic disease in pregnant patients, recent recommendations have widely expanded the use of thromboprophylaxis; the California Maternal Quality Care Collaborative’s Maternal Venous Thromboembolism Toolkit recommendations include significant expansions in the use of prophylactic and therapeutic heparins in the antepartum, peripartum, and postpartum periods.
Thrombophilia is emerging as an important cause of maternal thromboembolic risk. It is found in up to 17% of white North Americans. Thrombophilia can be divided into three general categories: familial, acquired, and mixed. Familial causes include protein C, protein S, and antithrombin III deficiencies, as well as factor V Leiden and prothrombin gene polymorphism. Conditions that increase the risk of thrombophilia include antiphospholipid antibodies and lupus anticoagulant. Mixed thrombophilia risk includes methylene tetrahydrofolate reductase polymorphism and hyperhomocysteinemia. Genetic thrombophilia is associated with other complications of pregnancy such as gestational hypertension, intrauterine growth restriction (IUGR), placental abruption, and stillbirth. In one study, 52% of women with severe gestational hypertension and associated problems also had thrombophilia. Preventing these thrombophilia-related side effects has proven to be difficult; antepartum dalteparin prophylaxis has shown no effect in preventing thromboembolic events, pregnancy loss, or placenta-mediated pregnancy complications in patients with thrombophilia in pregnancy.
Hypertensive Disorders of Pregnancy
Chronic hypertension complicates between 1% and 5% of pregnancies. It is defined as a blood pressure greater than 140/90 mmHg that either predates pregnancy or develops before 20 weeks of gestation. Gestational hypertension develops after 20 weeks of gestation and complicates between 5% and 10% of pregnancies. Gestational hypertension may reflect a familial predisposition to chronic hypertension or it may be an early manifestation of preeclampsia.
Preeclampsia is the most common disease that is unique to human gestation, occurring in 6% to 8% of pregnancies. In developed regions 12.9 % of deaths and 14% in developing regions were attributable to hypertensive disorders of pregnancy, most commonly preeclampsia. Diagnostic criteria are listed in Fig. 29.5 and include hypertension and symptoms of end-organ dysfunction.
The etiology is unknown, but the defining feature of preeclampsia is vasospasm affecting organs throughout the body. Cellular damage can be seen in endothelium, platelets, and trophoblasts. Vasoactive amines and prostaglandins are then released. Blood vessel narrowing and decreased elasticity are associated with increased vascular resistance. Women who become preeclamptic, unlike those with a normal pregnancy, do not lose sensitivity to angiotensin and catecholamines. Uterine Doppler studies show evidence of high vascular resistance and flow abnormalities.
An immune reaction involving trophoblast material and basement membrane, prostacyclin imbalance, and vascular nitric oxide dysfunction have all been suggested as important parts of the puzzle. A genetic analysis of its aspects is underway. General risk factors for the development of gestational hypertension can be found in Table 29.1.
As hypertensive disorders of pregnancy remain a significant cause of maternal morbidity and mortality, significant efforts have been made to improve, standardize, and expand appropriate and early care for gestational hypertension and preeclampsia. Core to these efforts are to standardize antepartum care and to monitor closely those at elevated risk of preeclampsia, early evaluation and treatment of all hypertension in pregnancy, early implementation of magnesium sulfate therapy for the treatment of preeclampsia, and improved education for patients and health-care professionals in both identification and treatment strategies.
Management of Maternal Hypertension and Preeclampsia
Obstetric maternal monitoring should be done in all women with gestational hypertension and preeclampsia. The aim in gestational hypertension is to watch for progression to preeclampsia and to treat as early as possible. In women with severe features of preeclampsia, the goal is to detect and avoid organ dysfunction.
Antepartum outpatient management of preeclampsia can be considered for select patients without severe features who have access to routine follow-up appointments, can adhere to their treatment plan, and have demonstrated fetal wellbeing and maternal stability. The fetus should be evaluated routinely (at least twice per week) by nonstress test or biophysical profile, and the mother should have routine blood pressure evaluation and laboratory studies, including complete blood count with platelet count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), and serum creatinine.
Management of patients with preeclampsia at term, preeclampsia with severe features, or patients with worsening disease demands immediate hospitalization. Patients with preeclampsia without severe features are routinely delivered at 37 weeks. Patients with severe features who have reached 34 weeks’ gestation are also routinely delivered; patients who are less than 34 weeks are typically monitored as inpatients with close observation, the goal being to postpone delivery until 34 weeks.
Magnesium Sulfate
The use of magnesium sulfate is considered a best practice for the control of seizures in preeclampsia and eclampsia. The MAGPIE Trial (Magnesium Sulfate for the Prevention of Eclampsia), a large, worldwide, placebo-controlled study, looked at this drug’s effectiveness (see Fig. 29.10 , later) The nearly 10,000 women in 33 countries who were recruited had exhibited at least two blood pressure readings of greater than 140/90 and had at least 1 + proteinuria. Of these, 4999 received an intravenous (IV) loading dose of 4 g and then maintenance dosages of 1 g/h intravenously or 5 g intramuscularly every 5 hours, and 4993 women received a placebo. There were fewer eclamptic seizures among women given magnesium sulfate than among those receiving placebo (40 [0.8%] vs 96 [1.9%]). Maternal mortality was lower among women given magnesium sulfate than among those given placebo (11 [0.2%] vs 20 [0.4%]). Of interest, one-third of the patients in this trial received nifedipine while receiving magnesium, and the rates of hypotension among these women were not higher than the rates seen in the placebo group.
Labetalol, Hydralazine, and Nifedipine
Hydralazine given intravenously had been recommended for over 30 years to control severe hypertension in pregnancy. However, in 11 trials involving 570 participants, hydralazine was compared with other antihypertensives for controlling severe hypertension in pregnancy. The authors found that parenteral hydralazine is not the drug of first choice for acute severe hypertension later in pregnancy because it is associated with more maternal and perinatal adverse effects than other drugs, particularly IV labetalol or oral or sublingual nifedipine. The trials compared IV hydralazine with other antihypertensives such as IV labetalol or oral sublingual nifedipine. Labetalol and nifedipine compared with hydralazine caused less hypotension, fewer cesarean sections, less placental abruption, and fewer low Apgar scores. Hydralazine was 3.2 times more likely to be associated with maternal hypotension than labetalol and nifedipine. Neonatal bradycardia was more common in the labetalol group, but only one neonate required treatment. A clinical advantage of nifedipine is that it is given by mouth; nursing staff may give it on an as-needed basis (every 30 minutes). Caution is advised because an interaction between calcium channel blockers and magnesium sulfate has been reported to produce profound maternal muscle weakness, as well as maternal hypotension and fetal distress.
Other Antihypertensives
Clonidine and prazosin have been used with good results for preeclampsia, but no large clinical trials are available. Sodium nitroprusside (SNP) infusions have been recommended for acute hypertensive crisis and to attenuate or treat hypertension associated with the induction and emergence of general anesthesia. It is rapid acting and somewhat unpredictable, usually requiring an intra-arterial catheter to avoid overshoot. Although effective, SNP is associated with cyanide toxicity in large doses.
Newer drugs may have a significant role to play in managing hypertensive emergencies in the pregnant patient. Nicardipine has been used as a replacement for SNP; it is a short-acting dihydropyridine calcium channel blocker given by IV infusion. Although there were originally concerns that adding nicardipine to magnesium may precipitate significant refractory hypotension, recent studies have not shown this to be the case. Clevidipine is a newer intravenous calcium channel blocker with similar action to nicardipine; however, specific studies in pregnancy are lacking.
Central Monitoring
Because the overwhelming majority of women with gestational hypertension have adequate left ventricular function and normal pulmonary artery pressures, clinicians must weigh the risk of central monitoring (central venous catheterization or pulmonary artery catheterization) against the usefulness of any additional information beyond that of standard monitoring and hemodynamic data. There are no large randomized trials showing improved maternal outcome with or without pulmonary artery monitoring in severe preeclampsia; in fact, with the increasing data calling into question the utility of pulmonary artery catheterization in the critical care population, as well as the growing adoption and acceptance of point-of-care ultrasound practices including bedside transthoracic echocardiography, the use of pulmonary arterial catheterization has decreased significantly. The American Society of Anesthesiologists (ASA) taskforce on obstetric anesthesia states: “It is not necessary to routinely use central invasive hemodynamic monitoring for severe preeclamptic parturients.”
Anesthetic Considerations in the Management of Hypertensive Disorders of Pregnancy
Old obstetric concerns that epidural and spinal analgesia in preeclampsia could lead to sudden maternal hypotension with maternal and fetal deterioration have not been borne out. Although no large randomized trial or meta-analysis addresses the effect of an anesthetic on maternal outcome in preeclampsia, several small trials, some prospective and randomized, are available for review. Thirteen trials involving 544 women with preeclampsia who received either regional or general anesthesia have been published. Three trials involved laboring preeclamptic patients. Ten trials involved preeclamptic patients who received a cesarean section, five of which compared regional anesthesia with general anesthesia. One trial compared spinal with epidural block in patients with severe preeclampsia. Two trials looked solely at spinal anesthesia and general anesthesia techniques. Overall, the incidence of maternal hypotension in these series was not considered to have an adverse clinical effect on either maternal or fetal outcome. In trials where general anesthesia was used, four trials found peri-induction hypertension difficult to control, whereas two found the incidence and severity of hypertension to be clinically acceptable. One series found the incidence of hypotension to be less in preeclamptic patients receiving spinal anesthesia than among healthy women.
Thrombocytopenia and Coagulopathy
Historically, a platelet count of less than 100,000/μL was considered an absolute contraindication to neuraxial anesthesia; this stance is no longer supported by the available science. The ASA Practice Guidelines now recommend that even the decision to order an intrapartum platelet count should be “individualized and based on a patient’s history… physical examination, and clinical signs.” No “safe” platelet count for neuraxial anesthesia has been identified by any professional organization; however, there is growing experience with thrombocytopenic patients with platelet counts down to 70,000/μL that imply safety. Experience below 70,000/μL is limited by a lack of data. The decision whether to place neuraxial anesthesia in the presence of thrombocytopenia should take into account not only the absolute platelet count, but the rate and direction of change and an assessment of platelet function.
Fluid Management and Pulmonary Edema
IV fluid management is part of standard anesthetic care. In patients with preeclampsia, this should be done with certain precautions. Decreased colloid oncotic pressure, increased hydrostatic pressure, and damaged capillary endothelium predispose to development of pulmonary edema. The incidence of pulmonary edema in pregnancy is 80–500 per 100,000 pregnancies. It is responsible for about 25% of pregnancy-related transfers to the intensive care unit. The causes of pregnancy-related pulmonary edema can be divided into tocolytic (36%), fluid overload (31%), gestational hypertension (26%), infection (4%), and other (3%). Approximately 3% of preeclamptic patients develop pulmonary edema, with 30% occurring before the birth and 70% occurring in the first 3 postpartum days. Maternal mortality may be as high as 10% and perinatal mortality as high as 50%. Those with fluid overload identified as the probable etiology had a significantly greater mean positive fluid balance (6022 ± 3340 mL).
The parturient is predisposed to pulmonary edema in that blood volume, hydrostatic pressure, and cardiac output increase, whereas colloid oncotic pressure decreases. In preeclampsia, pulmonary blood pressure and blood flow increase even further, colloid oncotic pressure decreases more, and capillary endothelium sustains damage. If lymphatic flow is impaired (e.g., by pulmonary tissue swelling and inflammation), the chances for symptomatic pulmonary edema increase still further. In a few cases, when preeclampsia has not been treated, cardiac failure exacerbates these conditions even further.
In patients with diastolic blood pressure greater than 100 mmHg, the recommendation is to perform the block accompanied by judicious use of vasopressors and intravenous fluid, as needed. It is usually possible to care for most of these patients without a central venous or pulmonary monitor. The use of supplemental intravenous 25% albumin, recommended in the past, is not widely used now. Decreasing vital capacity and a decrease in oxygen saturation by pulse oximetry may be indicative of developing pulmonary edema.
Cardiac Disease in Pregnancy
The differential diagnosis for sudden hemodynamic deterioration in a pregnant woman includes thromboembolic disease, hemorrhage, sepsis, preeclampsia, and cardiac disease. Maternal cardiac disease causes approximately 16% of all maternal deaths. The most common single cause of maternal mortality is cardiac disease. Cardiomyopathy, myocardial infarction, and aortic dissection are the most commonly reported conditions, with chronic and acquired diseases such as pulmonary hypertension, and valvular and congenital diseases being less common.
Preexisting cardiac disease carries significant risk as pregnancy progresses. The Cardiac Disease in Pregnancy (CARPREG) Studies ventured to quantify the risk of preexisting cardiac diseases by identifying predictors of adverse events in women with heart disease ( Fig. 29.6 ). A score, the CARPREG II score, was then derived from these findings by assigning a point score to 10 predictors and assessing risk of new cardiac events, including maternal cardiac death, cardiac arrest, sustained arrhythmia requiring treatment, left-sided heart failure, stroke or transient ischemic attack, cardiac thromboembolism, myocardial infarction, and vascular dissection. The CARPREG II scoring system and cardiac event risk with each point total is shown in Fig. 29.7 .
Cardiomyopathy
Peripartum cardiomyopathy occurs in 1 in 1500 to 1 in 4000 deliveries. There is no clear medical cause and onset can be 1 month prior to up to 5 months after delivery. Multiparity, older age, African heritage, and birth of twins appear to result in a higher incidence of cardiomyopathy. Authorities still do not agree on an etiology, although viral and immune responses have been implicated.
With peripartum cardiomyopathy there is usually no history of heart disease, and the onset of signs and symptoms is nonspecific and easily confused with normal changes of pregnancy, such as breathlessness, or with a viral chest infection. Easy fatigue, dyspnea, and ankle edema can be normal or abnormal. Therefore, clinicians should not ignore a patient who has dyspnea at rest or orthopnea, paroxysmal nocturnal dyspnea, chest pain, nocturnal cough, pulmonary rales, jugular venous distention, or regurgitant murmur. The electrocardiogram is often nondiagnostic, although findings may be consistent with cardiomegaly and include arrhythmias and ST changes. An echocardiogram remains the strongest diagnostic tool ( Box 29.1 ).
All four of the following:
- 1.
Heart failure within the last month of pregnancy or the first 5 months postpartum;
- 2.
Absence of prior heart disease;
- 3.
No determinable cause;
- 4.
Strict echocardiographic indication of left ventricular dysfunction:
- a.
Ejection fraction < 45% and/or fractional shortening < 30%
- a.
and
- b.
End-diastolic dimension > 2.7 cm/m 2
- b.
Patients with peripartum cardiomyopathy are best followed by a cardiologist familiar with the changes of pregnancy. Therapy includes diuretics, salt restriction, and volume overload avoidance. Vasodilation is aided with hydralazine, nitrates, or amlodipine. Angiotensin-converting enzyme (ACE) inhibitors are avoided because of reported teratogenic effects, fetal renal failure, and death. Amiodarone and verapamil are rarely used because of their potentially deleterious fetal effects. Most calcium channel blockers are associated with negative inotropy, but carvedilol may improve overall survival rate in parturient women with dilated cardiomyopathy, and perhaps in patients with peripartum cardiomyopathy. The use of intravenous immune globulin has been used successfully in some with peripartum cardiomyopathy. Patients with low ejection fractions (≤ 35%) may be given thromboprophylaxis with unfractionated or low-molecular-weight heparin.
Delivery of the fetus, if viable, improves cardiac function in most women with cardiomyopathy. Labor can be induced safely in most cases if cardiac stabilization can be achieved. Early consultation with an anesthesiologist is ideal. Less blood loss, lower infection rate, decreased surgical stress and occurrence of respiratory complications, and improved hemodynamic stability are advantages of vaginal delivery. The addition of a regional block decreases the variation in blood pressure, decreases the preload and afterload on the heart, and provides ideal conditions should instrumental delivery be required.
Cesarean delivery may be necessary if the mother is unstable or fetal indications are present. For cesarean section, either regional block or general anesthesia is an acceptable approach, depending on the patient’s coagulation status and ability to cooperate. In patients with symptomatic cardiomyopathy, direct arterial and central venous monitoring are advised. Some very ill patients who are gasping for breath may require induction in the sitting position, converting to supine once consciousness is lost. Etomidate for induction is advised if myocardial function is impaired. Some of these patients will require respiratory and inotropic support postoperatively and intensive medical care will be necessary. No large-scale studies on anesthesia outcomes exist, but descriptions of anesthesia with satisfactory outcome are available. In patients with severe myocardial impairment, a left ventricular assist device (LVAD) may be necessary, and consideration should be given to transplantation if ventricular function does not improve. Those who improve usually do so within 6 months after pregnancy. Mortality from peripartum cardiomyopathy ranges from 18% to 56%, and persistent cardiomegaly is associated with an 85% mortality.
Myocardial Infarction
Myocardial infarction is rare in this population; it occurs in about 1 in 10,000 deliveries. Risk factors include increased age, diabetes, smoking, hypertension, cardiomyopathy, and obesity. However, 78% have no apparent risk factors. Myocardial infarction is more common during the third trimester of pregnancy and mortality is high if it occurs within 2 weeks of delivery. Symptoms are classic, including anginal-type chest pain, diagnostic electrocardiographic changes, and increased cardiac-specific plasma enzymes. Labor may mask these symptoms and muscle creatinine kinase is normally elevated. Chest pain can also be caused by hemorrhage/shock, sickle cell crisis, severe preeclampsia, pulmonary embolism, and aortic dissection.
Immediate treatment may include oxygen therapy, sublingual nitroglycerin, aspirin, systemic heparinization, and coronary angiography with possible stenting. Tissue plasminogen activator (TPA) and assessment for emergency coronary artery bypass graft (CABG) have also been suggested. TPA and anticoagulants predispose to increased bleeding in the immediate postpartum period and are not advised during that period. These drugs also increase the risk of bleeding associated with regional block.
If possible, the use of ergot alkaloids, such as methylergonovine, should be avoided in these patients. Reports of peripartum myocardial infarction following ergot administration are frequent. If ischemia occurs, it is usually rapid and treatment with sublingual or intravenous nitroglycerin is recommended.
Aortic Dissection
Although rare, aortic dissection is associated with severe hypertension and inherited disorders such as Marfan syndrome. If symptoms are present, intrascapular or chest pain are most common. Increasing severity is associated with signs of cerebral, cardiac, renal, or limb ischemia and heart failure caused by aortic regurgitation or hemopericardium. Vaginal delivery with regional block and postpartum repair of descending aortic dissections is advocated if the fetus is viable. For ascending lesions, repair with the fetus in situ has been attempted for patients with fetuses before 28 weeks’ gestation. If the fetus is viable and because of the life-threatening nature of nonoperative treatment (80% mortality ), surgical repair and concomitant cesarean delivery should be anticipated.
Anesthetic goals include a reduction in pressure-flow gradient across the dissection. Epidural analgesia is suggested to reduce the wall tension, mean arterial blood pressure, and vessel shear stresses (cardiac output) seen in labor. For cesarean delivery, general anesthesia may be necessary if anticoagulation increases the risk of regional block. The risk of hypertension during induction of general anesthesia is well known and arterial and central venous lines are therefore recommended to provide beat-to-beat control of blood pressure and volume control as needed.
Chronic Cardiac Disease
As previously discussed, patients with significant prepregnancy cardiac disease are at much higher risk of peripartum cardiac events. Women with a history of cardiac disease in previous pregnancies or congenital heart disease with prior cardiac surgery can be identified, assessed, and counseled before becoming pregnant. In some cases, such as when severe pulmonary hypertension or Eisenmenger physiology is involved, pregnancy termination may be advised. Ready assessment and communication between cardiologist, obstetrician, and anesthesiologist should be the norm. The New York Heart Association functional class may not predict how the patient will adapt to pregnancy and regular echocardiographic assessment of the mother and fetal assessment are desirable.
Congenital Heart Disease
Because of improvements in diagnosis, treatment, and surgical therapy, women with repaired congenital heart disease (CHD) are reaching childbearing age in increasing numbers, and there is widespread encouragement in the literature for them to consider pregnancy. Pregnancy in women with CHD not complicated by Eisenmenger syndrome is associated with low mortality. However, potential risk factors for maternal morbidity include poor maternal functional class, poorly controlled arrhythmia, heart failure, cyanosis, significant left heart obstruction, and a history of cerebral ischemia. Cyanosis is a risk factor for fetal and neonatal complications. A second modified risk score has been developed, considering the unique challenges of both corrected and uncorrected congenital heart disease; risk of major pregnancy complications in these patients range from 2.9% to 70%, and the scoring system and breakdown are available in Fig. 29.8 . Medical or surgical termination of pregnancy in intermediate- or high-risk patients requires careful monitoring and should be done in a regional adult congenital heart disease center.
