The last 50 years have witnessed dramatic changes for the once threatened and limited life of the baby born with congenital heart disease. The advances in echocardiography, anesthesia, intensive care, and particularly cardiac surgery have facilitated the survival of babies born with even the most complex cardiac anomalies. Fifty years ago, only 25% of these infants would survive beyond the first year of life, but now more than 95% will survive to adulthood. This triumph of survival, which has evolved over the last few decades, has led to a “new population” of adults with congenital heart disease. This population size is estimated to be approximately 800, 000 in the U.S. (1).

Some patients may have mild defects and have never needed surgery; in others, the defect may have been missed and may not be discovered until adulthood. The majority, however, have had previous cardiac surgery and may consider themselves “cured.” The perception of “cure” is fostered by the surgical description “total correction,” which is applied to many operative repairs of complex congenital anomalies. In reality, there is almost no surgical cure for congenital heart disease, perhaps with the exception of a successfully ligated and divided ductus arteriosus. All other repaired lesions have the potential for residua and sequelae, and although this may be a painful realization for patients and their families, it is a fundamental and important concept. The misperception of “cure” has potentially serious consequences. Patients may forget to use antibiotic prophylaxis, are unlikely to endeavor to understand the nature of their anomaly, and, much more importantly, see no need to seek continued medical advice. As a result, residual lesions and sequelae are frequently overlooked until patients present with symptoms.

These patients are different from those with acquired heart disease. One of the most common presentations is with an arrhythmia, and cardiologists unfamiliar with congenital heart disease often focus on the electrophysiologic aspects of the symptom complex, unaware of the underlying hemodynamic problems so commonly associated with the onset of arrhythmias. In contrast to patients with acquired heart disease, who usually notice a distinct change in symptoms at the onset of their problems, patients with congenital heart disease, having lived with a lifelong cardiac problem, may not detect subtle changes in exercise capacity until they are significant. By the time the patients notice dyspnea and exercise limitation, valvular residua and ventricular dysfunction are often severe and irreversible.

All of these challenges emphasize the importance of impressing on patients, their families, and their physicians that all cardiac surgery is palliative rather than curative and that patients with congenital heart disease require lifelong follow-up at centers where expertise is available to deal with their complex problems. Two “simple” and two “complex” congenital cardiac anomalies that support this thesis will be reviewed.

Surgical repair of secundum ASD has been performed since 1954 and Murphy et al. (2) published the results of one of the early pioneer series of symptomatic patients. They were followed for 30 years, and patient survival was compared with age-matched control subjects (Figure 1). Although patients in the younger quartiles (<25 years) had normal survival, patients >25 years of age had a significantly reduced survival compared with control subjects, and this was even more striking in those >40 years of age. Late events, including atrial fibrillation, stroke, and heart failure were more common in patients repaired in adulthood, and this study was one of the first to demonstrate the benefits of early repair of secundum ASD in symptomatic patients. These residual problems after surgical repair in older patients have been confirmed in other studies (3- 5). Presumably, this relates to the long-standing deleterious effects of volume overload on the right-sided cardiac chambers, pulmonary hypertension, and right atrial enlargement that increase the vulnerability to atrial arrhythmias and stroke. In the series of Murphy et al. (2) 22% of late deaths were due to stroke. Thus, although this common and simple cardiac anomaly can be easily repaired surgically and, in recent years, closed percutaneously, it is not completely “cured” and is associated with ongoing morbidity and mortality.

Coarctation of the aorta is often regarded as a simple and isolated congenital anomaly, but is more correctly interpreted as part of a diffuse arteriopathy with a propensity to aneurysm formation and dissection remote from the coarctation site (6). In addition, more than 50% of patients have an associated bicuspid aortic valve (BAV) and 10% of patients have cerebral aneurysms demonstrated on magnetic resonance image scanning of the brain, suggesting a common etiologic mechanism (7). After successful coarctation repair, many patients consider themselves “cured,” and they are often discharged from physician follow-up. It has long been recognized that patients with repaired coarctation have premature morbidity and even mortality (8- 9). Late complications are common, and long-term survival may be far from normal. A detailed knowledge of these complex associations is critical in providing optimal care for these patients.

The vulnerability of these patients to premature cardiovascular complications was exemplified in the largest series ever reported after coarctation repair. Cohen et al. (8) reported the long-term follow-up of 646 patients operated on between 1946 and 1981. There were 17 perioperative deaths, and 58 were lost to follow-up. Of 571 survivors, 67 patients had 81 subsequent cardiac operations, most frequently for aortic valve replacement or re-coarctation. There were 87 late deaths, most commonly from acute myocardial infarction and sudden cardiac death, and the 30-year survival of this cohort was 72%. The mean age at death for this “simple” lesion was only 38 years. This emphasizes the continued morbidity from premature coronary disease and residual systemic hypertension, which is very common in such patients.

Other late cardiovascular complications requiring subsequent surgery are common. In two series from the Mayo Clinic (8,10), the most common reason for re-operation was aortic valve replacement, but mitral valve repair or replacement in addition to re-operation for aortic aneurysm, re-coarctation, and coronary artery bypass grafting are also represented. All of these congenital arterial abnormalities point toward a much more diffuse congenital cardiovascular problem than might be suspected from a superficial review of this “localized” narrowing in the aorta.

At 30-year follow-up, as many as 75% of patients have residual systemic hypertension (11). The pathophysiology of residual hypertension in such patients is poorly understood, and it occurs despite the absence of residual coarctation. Earlier reports suggested that hypertension was less likely to occur after early repair (12), but more recent reports suggest that even after early repair, residual hemodynamic abnormalities are common. O’Sullivan et al. (13) reported a 7- to 16-year follow-up of 119 children having coarctation repair at the age of 2 to 3 months. Of the patients, 28% had hypertension at rest or during ambulation, including 21% who had no residual aortic obstruction.

Even patients with normal blood pressure (BP) at rest often demonstrate an abnormal systolic response to physical exercise (14). This may contribute to increased left ventricular (LV) mass that has been reported even in normotensive patients after repair, and which, in and of itself, is another important predictor of cardiovascular complications (15). Whether this increased LV mass is a genetically determined hypertrophic response in coarctation or occurs because of persistent hemodynamic or hormonal abnormalities remains to be determined. Contributing causes may be changes in the arterial pressure wave propagation in the aorta and/or an increase in aortic stiffness or a “re-programming” of the sympathetic nervous and/or renin-angiotensin system (14).

Vascular reactivity and mechanical properties of large upper limb conduit arteries also continue to be impaired, even in normotensive adults when successful coarctation repair has been performed in the first few months of life (16). Studies examining the brachial artery response to flow-mediated dilation after reactive hyperemia (induced either by inflation and deflation of a pneumatic cuff around the forearm or administration of nitroglycerin) show less dilation than in control subjects (14). Whether this represents early “programming” of vascular reactivity in utero or an inherent arterial abnormality remains uncertain, but clearly this might contribute to persistent cardiovascular abnormalities such as increased LV mass and diastolic dysfunction (17).

This impaired arterial response to increased flow suggests an underlying endothelial dysfunction, and nitric oxide may play a role in this regard and contribute to hypertension. Nitric oxide is an important biological modulator that plays a significant role in vasodilation and BP control. Studies in animals have suggested a decreased nitric oxide bioavailability in the aortic segment proximal to the coarctation with up-regulation of endothelial nitric oxide synthase (eNOS) in this region. In contrast, distal to the coarctation, nitric oxide activity is normal (18). This suggests a local baromechanical or endothelial etiology rather than a circulating humoral factor that would be similar in both segments.

This “simple” focal lesion also posed residual problems related to other areas of the arterial system, which had a major impact on mortality and morbidity: stroke related to aneurysms of the Circle of Willis, aortic aneurysm and dissection related to the inherently abnormal aorta, and superimposed systemic hypertension (Figure 2). Ascending aortic aneurysm is the most frequently encountered complication. Aortic complications resulting in death or the need for surgery are frequent during adult life in patients with repaired coarctation. The co-existence of a BAV seems to influence their occurrence; in one series, the prevalence of aortic complications was 22% (29 of 134 patients) in those with a BAV compared with 8% (8 of 101 patients) in those without a bicuspid valve (19).

Bicuspid aortic valve occurs in more than one-half of the patients with coarctation and, therefore, periodic assessment of possible aortic valve disease is warranted even after successful coarctation repair. Bicuspid aortic valve is also one of the most common isolated congenital cardiac abnormalities, present in at least 1% of the population. Nitric oxide may also play a role in this pathophysiology, because mice lacking eNOS appear more likely to develop a BAV, although the mechanism remains uncertain. It is possible that valvular endothelium may generate signals that fine-tune the development of the primitive ventricular outflow tract to ensure the development of a normal tricuspid structure (20). Similarly, nitric oxide has been implicated in vascular remodeling in response to changes in luminal flow conditions; so lack of eNOS might also affect the modeling of the aortic isthmus during transition from the fetal to adult pattern of circulation. Whether a deficiency of the gene encoding eNOS exists remains unproven, but these abnormalities of nitric oxide in both BAV and coarctation are tantalizing and do implicate a unifying etiology. Further research in this area is necessary.

Bicuspid aortic valve is also an independent risk factor for progressive aortic dilation, aneurysm formation, and dissection. The aortic root dilation is unrelated to any hemodynamic disturbance of the valve itself, because >50% of young patients with a functionally normal aortic valve have echocardiographic evidence of aortic dilation (21- 22). This diagnosis, too, should prompt the beginning of lifelong surveillance of the arterial tree. Both coarctation and BAV are part of a spectrum of arteriopathy, and the search for a common pathophysiology is tantalizing. Two possible culprits include a genetic defect of fibrillin and/or an abnormality of nitric oxide metabolism.

Medial changes in the ascending aorta have been well demonstrated and are identical to the changes seen with coarctation. Histologic examination reveals smooth muscle cell loss, fragmentation of elastic fibers, and “pools” of basophilic ground substance in areas of cellular loss. The exact cause of these degenerative changes in the aorta is uncertain, but apoptosis may be an important underlying mechanism causing the smooth muscle cell loss (23). There may be an underlying genetic stimulus for programmed cell death in the aortic media, and the recent observation that BAV is heritable supports this concept (24).

Interestingly, de Sa et al. (25) reported that degenerative changes in the media of the pulmonary artery are also more common in patients with BAV, perhaps because the aorta and pulmonary artery share the same embryologic origin from the conotruncus. This would also support the concept of a more diffuse “arteriopathy” and, perhaps, that BAV and coarctation are part of a continuum of arterial abnormality.

Decreased levels of fibrillin-1 have also been reported in the ascending aorta of patients with BAV, and this deficiency of fibrillin-1 may cause smooth muscle cells to detach from the elastic laminae with the release of matrix metalloproteinases. These matrix metalloproteinases weaken the aortic wall, degrade the fibrillin-1, and contribute to the aortic dilation (26). Although the gene for fibrillin-1 may be structurally normal in patients with BAV, it is possible that transcriptional elements that control protein production may be defective, thereby precipitating the arterial changes.

These important residua and sequelae support the notion that patients with repaired coarctation need lifelong follow-up with periodic imaging of the aortic valve and the entire aorta. Two-dimensional echocardiography is particularly helpful in this regard, and echo-Doppler assessment of the aorta facilitates measurement of the aorta, coarctation site, and aortic valve. Magnetic resonance imaging provides a complementary adjunct to the evaluation and allows visualization of the entire aorta as well as facilitating the detection of possible aneurysm formation at the site of previous repair (usually not seen by echocardiography or chest radiograph). Histologic abnormalities of the ascending aorta are common: the so-called “cystic medial necrosis,” in which there is fragmentation of the elastic fibers and accumulation of ground substance, making the aorta vulnerable to aneurysm formation and dissection.

Blood pressure control should be meticulous to minimize shear stresses on the aorta; this should be evaluated both at rest and on exercise. Beta-blockade is ideal therapy in this setting, although, as yet, there is no definite evidence that they prevent aortic dilation as demonstrated in the Marfan syndrome (27). Ambulatory BP monitoring may be helpful if there is doubt about the effectiveness of BP control (13). Careful attention should also be paid to BP at the time of pregnancy, because hormonal changes exacerbate the distensibility of connective tissues, including those in the aorta, and dissection and rupture of the aorta may occur during pregnancy, especially around the time of delivery.

Tetralogy of Fallot is one of the most common cyanotic defects encountered in infancy. Lillehei et al. (28) performed the first repair in 1954, and operative results from even the first 106 patients from the “pioneer era” were excellent with a 30-year survival of 91%. The original repair involved closure of the ventricular septal defect and resection of muscle from the infundibular region of the right ventricular (RV) outflow tract (29). Because of the common finding of a small annulus, however, subsequent repairs often involved either a patch across the RV outflow or the placement of a transannular patch. Although the patch is effective in relieving the obstruction, it distorts the pulmonary valve apparatus and pulmonary regurgitation inevitably results.

Pulmonary regurgitation is tolerated well for many years, even for decades, but the chronic effects of long-term volume overload of pulmonary regurgitation eventually has a deleterious effect on RV function. Exercise capacity declines, secondary tricuspid regurgitation may occur, and supraventricular and ventricular arrhythmias may supervene. Sudden death may be the presenting feature. Patients often do not notice symptoms until RV dysfunction is severe. There is also a “ventricular-ventricular interaction,” which occurs in the setting of RV enlargement and systolic dysfunction, and concomitant LV dysfunction is often observed (30). The mechanism that links RV dysfunction to LV dysfunction, however, is incompletely understood.

Pulmonary regurgitation is frequently overlooked on physical examination, since the diastolic murmur is soft and short because there is rapid equalization of the diastolic pressures in the pulmonary artery and RV. The regurgitant jet is also frequently missed on two-dimensional echocardiography, because the jet has a low velocity and the flow is laminar.

Any patient with previous surgical repair of tetralogy of Fallot should have a normal heart size on chest radiograph, and the observation of an increased cardio-thoracic ratio should prompt a thorough search for a residual hemodynamic lesion. This may be a residual ventricular septal defect, aortic regurgitation, or pulmonary stenosis. The most common problem, however, is pulmonary regurgitation, and pulmonary valve replacement is the most common indication for re-operation late after tetralogy repair (31). Absence of symptoms does not reflect the functional derangement of the RV after tetralogy repair. Exercise testing has demonstrated compromised exercise performance, and hemodynamic data show elevated RV end diastolic volumes along with reduced cardiac output and ejection fraction in patients with pulmonary regurgitation (32- 33).

Pulmonary valve replacement can be accomplished with a low surgical risk (1% to 2%) in experienced centers, but should be performed before there is irreversible RV dysfunction and the increased propensity for ventricular arrhythmias and sudden cardiac death (34- 35). Timing of pulmonary valve insertion is critical in reducing RV size, preserving myocardial function, and preventing the development of arrhythmias (35). The precise indication for pulmonary valve replacement remains uncertain, although mounting evidence suggests that in many centers it has been performed too late. Even when subjective improvement in clinical symptoms is noted there is often no improvement in either RV volumes or function (36). This must be balanced against the concern of replacing the pulmonary valve early, however, which means the patient faces another re-operation approximately 10 years later owing to structural degeneration of the tissue prosthesis. Patients who require a pulmonary valve replacement in the third or fourth decade may require several re-operations in their lifetime.

Arrhythmias are also common sequelae of repaired tetralogy. Atrial arrhythmias may be present in approximately one-third of patients and are a major source of morbidity (37). Although the prevalence of sustained ventricular tachycardia is low, it is believed to be responsible for the small but definite incidence of sudden cardiac death in postoperative patients. Pulmonary regurgitation is the predominant underlying hemodynamic problem. Potential variables predictive of death include older age at repair, heart failure, residual or recurrent ventricular septal defect, and elevated RV pressure. Ventricular ectopy is common in this patient population, but neither frequent ectopy nor non-sustained ventricular tachycardia on ambulatory Holter monitor reliably identifies patients at risk of ventricular tachycardia or sudden death. Right bundle branch block is the expected pattern on the electrocardiogram in 95% of patients, and a “mechano-electrical” association has been demonstrated, showing that the larger the RV size, the greater the trend toward a longer electrocardiographic wave (QRS) duration. Mean QRS duration tends to be longer (>180 ms) in populations of patients who die suddenly than in their healthy counterparts (38). The rate of late progression of QRS duration may also be a warning sign (39). Nonetheless, these findings, although having significance in large populations of patients, have limited prognostic value in an individual patient and have poor sensitivity and specificity.

The issue of whether electrophysiology studies are useful to predict the development of clinical ventricular tachycardia has also not been resolved. Failure to induce ventricular tachycardia may be interpreted as a favorable prognostic sign, but a high proportion of patients have false-negative studies despite aggressive stimulation protocols (40). It has been suggested that the inclusion of polymorphic ventricular tachycardia in the definition of inducibility improves the sensitivity but with a slight reduction in specificity (41). Furthermore, no study has conclusively demonstrated that anti-arrhythmic medication improves survival. Thus, risk stratification in postoperative patients after tetralogy repair remains a major challenge, and the indications for implantation of a defibrillator device remain imprecise. Targeted arrhythmia procedures at the time of pulmonary valve replacement utilizing intraoperative electrophysiological mapping and/or cryoablation appear to decrease the incidence of arrhythmias, at least in the short term (42).

Thus, since the time of the first surgical repair of tetralogy of Fallot, the operation has been referred to as “total correction.” This misnomer fosters the erroneous belief held by patients that they are “cured” and that no further surgical intervention will ever be necessary. This may prevent them from seeking regular cardiac follow-up, so re-operation to replace the pulmonary valve is often performed later than ideal, when RV dysfunction is irreversible and patients are vulnerable to arrhythmias and sudden cardiac death. Resentment is common when patients learn, to their dismay, that another operation is necessary.

Ventricular dysfunction is a common problem in many adults with congenital heart disease and is particularly common in those born with transposition of the great arteries. Simple transposition (d-transposition) is the most common cyanotic abnormality in newborns, in which the right atrium is connected to the RV (atrioventricular [AV] concordance) which, in turn, gives rise to the aorta (ventriculo-arterial discordance). The left atrium enters the LV, which gives rise to the pulmonary artery. Since the 1980s, this defect has usually been repaired in infancy with an arterial switch procedure so that the LV is restored to function as the systemic ventricle. In the 1960s, however, the only reparative procedures were the Mustard or Senning operations that redirect the blood flow via an atrial baffle so that venous blood is directed into the LV and then to the pulmonary artery, and pulmonary venous blood is directed to the RV and thence to the aorta. These operations resulted in a dramatic improvement in the lives of cyanotic infants, who were rendered pink and healthy, most of whom continue to do well three decades later. Long-term problems are inevitable, however, particularly as the morphologic RV continues to function as the systemic ventricle This may function well for decades (43), but function deteriorates eventually, often with concomitant tricuspid regurgitation (44- 45). Pulmonary artery banding to “retrain” the LV followed by an arterial switch operation has provided improvement in systemic ventricular function in some young patients and adolescents, but results of LV “conditioning” have been less consistent in adults (46). As a result, this approach has been largely abandoned in adults in favor of cardiac transplantation.

A similar situation exists in those patients with congenitally corrected transposition who have AV and ventriculo-arterial discordance; congestive heart failure is common by the fourth and fifth decades (47). In this situation, the morphologic RV is also the systemic pump, and the systemic AV valve (the tricuspid valve) is frequently congenitally abnormal and vulnerable to regurgitation. The diagnosis of congenitally corrected transposition may be missed by the unwary physician who may overlook the abnormal position and appearance of the ventricles on echocardiography and fail to observe the more inferior tricuspid valve (closer to the cardiac apex) on the left side. The tricuspid valve always enters a morphologic RV, and the leaflets insert directly into the ventricles rather than having attachments to papillary muscles. In one study from the Mayo Clinic (48), the diagnosis of congenitally corrected transposition had been missed in 7 of 44 adults who had had a cardiology consultation, despite the performance of one or more imaging studies (echocardiography and/or cardiac catheterization). More importantly, most of these patients were referred to a tertiary care center late for an evaluation. Of 30 patients who needed a systemic AV valve replacement for severe systemic AV valve regurgitation, 16 had >3/4 grade AV valve regurgitation and had established systemic ventricular dysfunction for more than six months before referral. Even though AV valve replacement can be accomplished at a relatively low risk in experienced hands, ventricular dysfunction has important prognostic implications. In one surgical series of 40 adult patients reported by van Son et al. (49), the mean preoperative ejection fraction was 48% (range 20% to 60%). There were four patients who died perioperatively, and in a mean follow-up of 4.6 years, eight patients subsequently died. The cause of death in all 12 patients was ventricular failure. Survivorship correlated not only with more recent operations as surgical techniques improved, but more importantly, with pre-operative ejection fraction (ejection fraction >40%).

Therefore, it is imperative that patients with congenitally corrected transposition have a detailed evaluation on a regular basis. This should include a physical examination, chest X-ray, electrocardiogram, and echocardiogram performed by someone with expertise in imaging complex congenital anomalies. An increasing cardiothoracic ratio or deterioration of systemic ventricular function should prompt a careful search for systemic AV valve regurgitation. Implantation of an endocardial pacemaker may also warrant more frequent follow-up, because the change in septal activation causes septal “shift” and secondary dilatation of the systemic AV annulus and subsequent regurgitation (Figure 3).

It is tempting to extrapolate the same treatment strategies for those with ventricular dysfunction with acquired heart disease and those with congenital heart disease and systemic (morphologic right) ventricular dysfunction. To date, however, there is very little evidence that administration of angiotensin-converting enzyme inhibitors or beta-blockers has a beneficial effect on exercise capacity, ejection fraction, or length of life in these patients (50). Indeed, some patients may have abnormalities of ventricular filling related to abnormal flow through the atrial baffle. Patients with postoperative transposition and those with congenitally corrected transposition are particularly vulnerable to sinus node disease, and injudicious use of beta-blockers may precipitate more profound conduction disturbances and complete heart block. Multi-center trials are needed to determine the appropriate therapeutic modalities for patients with congenital heart disease and systemic ventricular dysfunction when the RV functions as the systemic ventricle.

All of the lesions discussed exemplify the notion that babies born with congenital heart disease are seldom “cured.” Despite the phenomenal advances in diagnostic, medical, and operative care that have occurred in the last 50 years, many circulations still do not function normally, and residual problems are common. Cardiologists throughout the U.S. still have little opportunity for exposure to adult congenital heart disease, and despite training recommendations, few trainees have the opportunity to see such patients during their fellowship. Many cardiologists, therefore, have little understanding about the complexities of many postoperative residua and sequelae.

Patients often have a poor understanding of their anatomy and physiology and little concept of the possible challenges the future may hold. The surgical descriptor “total correction” fosters this misunderstanding, because patients see little need to understand what was “fixed” and no incentive to seek medical follow-up. Even when they do seek follow-up, insurance obstacles often obstruct the path to the few tertiary care centers where interdisciplinary expertise and resources are available. All patients confront many ongoing cardiac issues related to their congenital heart disease, not to mention the impact that acquired hypertension, coronary artery disease, and other cardiovascular problems may impose on their underlying cardiac anomaly. This becomes an increasing problem as the population ages (Figure 4). In those who have had previous surgery, many face multiple re-operations in their lifetime (Figure 5), and each subsequent operation poses an increasingly higher risk.

In conclusion, are these patients “born to be bad”? In many ways, the answer is yes. They are seldom “cured” by surgery and continue to have cardiac problems. Much time, money, and effort has been devoted to secure their survival, and unfortunately, very little thought has been given to providing for their long-term care. These survivors are extraordinarily courageous and usually, determined to work, contribute to society, and be as normal as possible. In adulthood, they often receive no care or suboptimal care, perhaps the worst of any cardiovascular subspecialty. The cardiology community serves them poorly, and, as we look to the future, we must make provision for lifelong care by trained physicians with expertise in their complex problems.