Author + information
- Received June 8, 1998
- Revision received February 17, 1999
- Accepted March 15, 1999
- Published online July 1, 1999.
- Koichiro Niwa, MD, FACCa,b,c,d,
- Joseph K Perloff, MD, FACCa,b,c,d,* (, )
- Samuel Kaplan, MD, FACCa,b,c,d,
- John S Child, MD, FACCa,b,c,d and
- Pamela D Miner, MN, NPa,b,c,d
- ↵*Reprint requests and correspondence: Dr. Joseph K. Perloff, Division of Cardiology, Room 47-123 CHS, UCLA Center for the Health Sciences, Los Angeles, California 90095-1679
Morbidity and mortality patterns were characterized in adults with the Eisenmenger syndrome when two ventricles with a ventricular septal defect (VSD) joined two great arteries or one great artery, or when one ventricle joined two great arteries.
Although afterload in these disorders differs, clinical differences have not been defined.
Seventy-seven patients were studied. Group A comprised 47 patients with VSD, aged 23 to 69 years (mean 39.5 ± 10.2), follow-up 5 to 18 years (mean 7.2 ± 4.9); group B, 14 patients with truncus arteriosus, aged 27 to 50 years (mean 33.7 ± 7.3), follow-up 6 to 18 years (mean 7.7 ± 5.1), and group C, 16 patients with univentricular heart, aged 18 to 44 years (mean 30.6 ± 8.4), follow-up 5 to 15 years (mean 4.4 ± 4.2). Echocardiography established the diagnoses and anatomic and hemodynamic features. Data were compiled on tachyarrhythmias, pregnancy, infective endocarditis, noncardiac surgery and the multisystem disorders of cyanotic adults.
Thirty-five percent of the patients died. Sixty-three percent of deaths were sudden, and resulted from intrapulmonary hemorrhage, rupture of either the pulmonary trunk, ascending aorta or a bronchial artery, or vasospastic cerebral infarction, or the cause was unestablished. There were no documented tachyarrhythmic sudden deaths.
Medical management of coexisting cardiac disease, multisystem systemic disorders, noncardiac surgery and pregnancy has reduced morbidity. Increased longevity exposed patients to proximal pulmonary arterial aneurysms, thromboses and calcification; to truncal valve stenosis and regurgitation; to semilunar and atrioventricular valve regurgitation, and to major risks of nontachyarrhythmic sudden death.
In his landmark article, Paul Wood defined the Eisenmenger syndrome as “pulmonary hypertension with reversed central shunt (1).” Nonrestrictive ventricular septal defect (VSD), truncus arteriosus (TA) and univentricular heart (UVH) were considered physiologically if not clinically indistinguishable because the ventricle(s) ejected at but not above systemic arterial pressure (1). Current reports still argue that prognosis is not influenced by the location of the intracardiac defect (2). We characterized morbidity and mortality patterns when two ventricles with a nonrestrictive VSD communicate with either two great arteries or one great artery, or when one ventricle communicates with two great arteries.
The Ahmanson Adult Congenital Heart Disease Center at the University of California at Los Angeles is a tertiary care referral facility. The Registry consists of outpatients and inpatients who either are referred from our Pediatric Cardiology Division upon reaching age 18 years, or are referred from physicians in the Greater Los Angeles area, the state of California, the U.S. and abroad. Patients are seen initially and then in follow-up by the professional staff of the Center, who are solely responsible for entering clinical data into research files that are maintained separately from hospital charts and records. Data are therefore not derived from retrospective review of hospital charts and records or from questionnaires.
“Two ventricles” refers to normally formed noninverted chambers with inlet and outlet portions divided by a septum (3). “Truncus arteriosus” refers to a single great artery that leaves the base of the heart, gives rise to the coronary, pulmonary and systemic arteries, and has a single semilunar valve (3). Type 1 is represented by a short main pulmonary artery that arises from the truncus, and type 2 by pulmonary arterial branches that arise by separate ostia (3). “Univentricular heart” applies when one ventricular chamber qualifies as a left, right or indeterminate ventricle on purely morphologic grounds, and receives the entire flow from right and left atria, both of which, with the atrioventricular junction, are related to the single ventricle (3). Univentricular heart as just defined does not include tricuspid or mitral atresia (3).
Diagnoses were established by transthoracic echocardiography in all patients. Color flow imaging with spectral Doppler quantified valvular regurgitation (4). Transesophageal echocardiography (4)was used in three patients with UVH to clarify anatomic and hemodynamic details. Sixty-seven patients had been catheterized long before referral; reliable data were available in 39 of these patients. We did not catheterize any patient.
Of 1,065 referrals during 1978–1995, 200 of 245 cyanotic patients had pulmonary vascular disease (PVD) on initial visit. Of these 200 patients, 138 were unoperated; two had interventions that left no detectable residua. Of these 140 patients (138 plus two), 77 (55%) met inclusion criteria of PVD with either nonrestrictive VSD (group A), 47 patients, aged 23 to 69 years, mean 39.5 ± 10.2, 19 male and 28 female, follow-up 5 to 18 years, mean 7.2 ± 4.9; TA (group B), 14 patients, aged 27 to 50 years, mean 33.7 ± 7.3, five male and nine female, follow-up 6 to 18 years, mean 7.7 ± 5.1 years, or UVH (group C), 16 patients, aged 18 to 44 years, mean 30.8 ± 8.4, nine male and seven female, follow-up 5 to 15 years, mean 4.4 ± 4.2. Five patients in group A, none in group B and two in group C were lost to follow-up.
In group A, the VSD was perimembranous in 24 patients, inlet septum in 15 (three with Down syndrome), outlet septum in 5 and muscular trabecular septum in 3. Coexisting congenital malformations included a restrictive interatrial communication in three patients, a restrictive patent ductus in two and dextrocardia in situs inversus in one. In group B, 12 patients had TA type 1, and two had TA type 2. In group C, 14 patients had a UVH of left ventricular morphology and two of right ventricular morphology.
In all patients, information was gathered on semilunar and atrioventricular valve regurgitation, ventricular function (transthoracic echocardiography), atrial and ventricular arrhythmias, pregnancy, infective endocarditis and noncardiac surgery. In groups A and B, right and left ventricular ejection fractions were derived from visual estimates utilizing transthoracic echocardiography. In group B, left ventricular ejection fractions were quantified in the four patients with depressed function. In group C, all ejection fractions were quantified (4). Functional class was based on the University of California at Los Angeles Classification for congenital heart disease (5). Event recorders or 24-h Holter monitors were obtained in patients with palpitations or when atrial or ventricular arrhythmias were detected on 12-lead scalar electrocardiograms with rhythm strips.
Hematologic data in all patients included hematocrit (automated particle counter), hemoglobin, iron indexes, automated platelet counts, transcutaneous fingertip systemic arterial oxygen saturation, hyperviscosity symptoms (based on a standard set of questions) (5)and the incidence of phlebotomy. A bleeding tendency was defined as cutaneous bruising, epistaxes and gingival bleeding (5). Calcium bilirubinate gallstones were identified at laparotomy or when clinical suspicion warranted abdominal ultrasound. Renal involvement in all patients was based on proteinuria, blood urea nitrogen, serum creatinine and uric acid levels. Abnormalities of the pulmonary circulation, apart from PVD, were represented by pulmonary hemorrhage (external and/or intrapulmonary), and in situ thromboses, aneurysms and calcification of hypertensive proximal pulmonary arteries identified by computerized tomography (CT). Long bone involvement was expressed as symptomatic hypertrophic osteoarthropathy confirmed by X ray. Asymptomatic hypertrophic osteoarthropathy was not routinely sought. Cerebral disorders included syncope, brain abscess and stroke.
Seventeen of the 18 necropsied patients were studied by the same cardiac pathologist.
Results are expressed as mean ± standard deviation unless otherwise indicated. The Kaplan-Meier method was used to establish three to five year actuarial survival rates after initial visit. Differences between curves were analyzed by log-rank (Mantel-Cox) statistic (Stat View [version 5.0] for Macintosh, Los Angeles, California).
Systemic ventricular (i.e., left ventricular or univentricular) ejection fractions in groups A, B and C were compared by dividing each of these three groups into two different subgroups: normal ejection fraction (group A, 47 patients; group B, 9 patients; group C, 0 patients) and abnormal ejection fraction (group A, 0 patients; group B, 5; group C, 16 patients). Because 20% of cells had expected counts <5, we used Fisher exact test to compare proportions (JMP software, SAS Institute, Cary, North Carolina).
Group A: ventricular septal defect (47 patients)
Six patients (13%) had been hospitalized for congestive heart failure in the first few years of life. In six patients (13%) cyanosis dated from childhood but not from birth. Systemic arterial O2saturations on initial visit and selectively on revisits were 75% to 84% (mean 79%).
Because the VSDs were nonrestrictive, systolic pressure in the pulmonary trunk and right ventricle was assumed to be the same as in aorta and left ventricle, an assumption confirmed in 39 patients by continuous wave Doppler velocities of tricuspid regurgitation jets (Table 1). Because the shunts were all right to left, pulmonary vascular resistance was assumed to exceed systemic.
Twelve pregnancies were experienced by seven of 28 women. There were seven live births, three spontaneous abortions, one induced abortion and one preterm stillbirth. Four deliveries were vaginal and three were cesarean for obstetrical indications. Five offspring were dysmature, two of whom had restrictive VSDs that closed in the first year of life. Gestational complications included postpartum syncope, hemorrhage (vaginal delivery) and sustained postpartum symptomatic deterioration. There were no maternal deaths. None of the 19 men fathered offspring.
There were no cases of infective endocarditis.
Eight patients (17%) had calcium bilirubinate gallstones, four identified at laparotomy, and four by abdominal ultrasound.
Gouty arthritis occurred in 11 patients (23%).
Hemoptysis recurred in 27 patients (57%), and was severe in 17 (36%). Computerized tomography scans disclosed intrapulmonary hemorrhage in 7/27 patients (Fig 2a). Computerized tomography scans of proximal pulmonary arteries in 18 patients identified large in situ thromboses (Fig. 3a)with moderate to marked mural calcification in six patients, and aneurysmal dilation of the pulmonary trunk and proximal branches in two patients. Proximal pulmonary arterial thromboses embolized in two patients, causing hemorrhagic pulmonary infarction.
Four patients (8%) had symptomatic hypertrophic osteoarthropathy.
One patient had two remote brain abscesses without neurologic residua. None experienced syncope or stroke.
The ages of 33 patients living at the time of study or when lost to follow-up (five patients) ranged from 23 to 52 years (mean 37 ± 8.1 years). Functional class was II or III (5). Twenty-nine patients (89%) were working. Ages of 14 deceased patients were 26 to 69 years (mean 45 ± 5 years). Ten of the 14 (71%) died suddenly, were functional class II or III and were working immediately before death. Nine of the 14 deceased patients underwent necropsy which disclosed the cause of sudden death as massive intrapulmonary hemorrhage in two (Fig. 2b), rupture of an aneurysmal pulmonary trunk in one (Fig. 4b), dissection of the ascending aorta in one, and unestablished in three patients. One other necropsy patient died of a vasospastic cerebral infarct caused by intractable migraine headache. Additional cardiac necropsy findings included massive in situ thromboses in dilated proximal pulmonary arteries (Fig. 3b)and an unruptured aneurysm of the pulmonary trunk (Fig. 4a). Of five deceased patients who were not necropsied, death was sudden and unexplained in three, due to right ventricular failure in a fourth and due to legionnaires’ pneumonia in the fifth, a 69-year-old man who was the oldest survivor.
Group B: truncus arteriosus (14 patients)
No patient had been hospitalized for congestive heart failure in early childhood. Neonatal cyanosis was reported in five patients (36%). Eleven (79%) were moderately to severely cyanotic when first seen. Systemic arterial O2saturations on initial visit and selectively on revisits were 76% to 85% (mean 80%).
Because the VSDs were nonrestrictive, systolic pressure in the right and left ventricles and in the truncus and pulmonary arteries was assumed to be the same except in three patients with truncal valve stenosis (Table 1). These assumptions were confirmed by continuous wave Doppler velocities of tricuspid regurgitation jets in eight patients. Pulmonary vascular resistance was assumed to equal or exceed systemic because of the degree of systemic arterial hypoxemia.
One pregnancy among nine women was terminated by induced abortion. None of the five men fathered offspring.
There were no cases of infective endocarditis.
Gallstones, either occult or overt, were not identified.
Gouty arthritis occurred in two patients (14%).
Hemoptysis recurred in three patients (21%). Computerized tomography scans disclosed intrapulmonary hemorrhage in two patients. Intrapulmonary embolization from a proximal pulmonary arterial thrombus in one patient caused hemorrhagic pulmonary infarction. Mural calcification was absent to mild.
One patient had symptomatic hypertrophic osteoarthropathy.
Remote brain abscess occurred in one patient. None experienced syncope or stroke.
ages of 10 patients living at the time of study were 27 to 42 years (mean 31 ± 6.3). None was lost to follow-up. Functional class was II or III, and all were working. Four deceased patients were aged 31 to 50 years (mean 41.5 ± 1). Three of the four died suddenly, were functional class II or III and were working immediately before death. Necropsies in the three sudden death patients disclosed thromboses in dilated proximal pulmonary arteries, a calcified stenotic truncal valve and aneurysms of proximal pulmonary arteries. One patient died of right ventricular failure initiated by atrial flutter.
Group C: univentricular heart (16 patients)
Four patients with UVH of left ventricular morphology and two of right ventricular morphology had been hospitalized for congestive failure in early childhood. Cyanosis dated from birth in five patients (31%). Eleven (69%) had moderate to severe cyanosis when first seen. Systemic arterial O2saturations on initial visit and selectively on revisits were 73% to 85% (mean 81%).
Because there was neither pulmonary stenosis nor stenosis of the outlet foramen (morphologic left ventricle), systolic pressure in the UVH, aorta and pulmonary trunk was assumed to be the same (Table 1). Because of the degree of systemic arterial hypoxemia, pulmonary vascular resistance was assumed to equal or exceed systemic.
Three pregnancies in three of seven women culminated in two spontaneous abortions and one induced abortion without maternal morbidity. None of the nine men fathered offspring.
Infective endocarditis occurred in two patients. The sites were a malformed incompetent left atrioventricular valve and a malformed incompetent right atrioventricular valve with septic pulmonary embolization.
Calcium bilirubinate gallstones were identified at laparotomy in one patient.
Gouty arthritis occurred in four patients (25%).
Hemoptysis recurred in six patients (38%). Computerized tomography scans disclosed intrapulmonary hemorrhage in three patients. One patient had large in situ thromboses in dilated proximal pulmonary arteries, the walls of which were moderately calcified.
Two patients (12%) had symptomatic hypertrophic osteoarthropathy.
No patient experienced syncope, stroke or brain abscess.
The ages of seven patients living at the time of study or when lost to follow-up (two patients) were 18 to 38 years (mean 33 ± 6). Functional class was II or III. Five of the seven were working. Nine deceased patients were aged 18 to 44 years (mean 31 ± 6.6). Four of the nine died suddenly. Sudden death in two was due to massive intrapulmonary hemorrhage caused by a ruptured bronchial artery, and to cardiac arrest during ventricular failure initiated by atrial fibrillation. The cause of sudden death was not established in the other two patients. Three of the four sudden death patients were functional class II or III and were working immediately before death. The other three had massive in situ thromboses in dilated proximal pulmonary arteries. Eight of the nine deceased patients had UVH of left ventricular morphology, ejection fractions <30%, moderate to severe pulmonary valve and atrioventricular valve regurgitation and mild to moderate aortic regurgitation.
The three- to five-year actuarial survival rates after initial visit were significantly different among groups A, B and C (log-rank statistics: chi-square = 8.099, df = 2, p = 0.0174) (Fig. 5).
The proportions of patients having abnormal systemic ventricular (left ventricular or univentricular) ejection fractions were significantly different among groups A, B and C (results of Fisher exact test: group A vs. group B, p < 0.001; group A vs. group C, p < 0.001; group B vs. group C, p < 0.001).
We characterized morbidity and mortality patterns in the Eisenmenger syndrome represented by nonrestrictive VSD, TA or UVH. In each malformation, the ventricle(s) eject at but not above systemic arterial pressure, but afterload defined as systolic tension in ventricular walls—that is, the load that the myocardium must bear to contract—is related not only to systemic arterial pressure, but also to ventricular volume, mass and geometry, and to great arterial compliance. These variables differ in the three malformations.
Early childhood morbidity expressed as congestive failure occurred in 13% of group A, none in group B and 38% in group C (Table 1). Persistent neonatal elevation or an early rise in pulmonary vascular resistance apparently curtailed the left to right shunt and improved the congested circulatory state (3).
Systemic ventricular function (left ventricular or univentricular) was better when two ventricles communicated with either two great arteries (VSD Eisenmenger) or one great artery (TA), than was function when a UVH communicated with two great arteries (Table 1). Right ventricular wall motion (normal or mildly to moderately depressed) was similar in groups A and B (Table 1). These observations underscore the importance of ventricular–ventricular interaction attributed to mechanical coupling through the ventricular septum and through the free walls of right and left ventricles that are in syncytial muscle fiber continuity (6,7). Univentricular heart of left ventricular morphology functions better than UVH of right ventricular morphology (8)(Table 1). Inadequate physiologic adaptation in the latter adversely affects wall tension and ejection fraction (8).
In group B, the incidence of moderate to severe truncal valve regurgitation or stenosis was 43% and 21% respectively (Table 1). Because a truncal valve is biventricular, these acquired loading conditions necessarily affected both ventricles.
Afterload reduction, the safety and efficacy of which have been questioned in the Eisenmenger syndrome, was not used in our patients. Angiotensin-converting enzyme inhibitors have been used selectively, however, with reported success (9).
Supraventricular tachyarrhythmias occurred in 13%, 36% and 25% of groups A, B and C respectively (Fig. 1). Atrial tachyarrhythmias exerted the most adverse effects on UVH. Nonsustained monomorphic ventricular tachycardia occurred in six group A patients and in one group C patient. Two group A patients had nontachycardiac monomorphic ventricular ectopic rhythms terminated by cardioversion.
Pregnancy incurred significant morbidity, underscoring maternal risks of PVD per se, in addition to risks of fetal wastage and dysmaturity associated with cyanosis (2,10). Eleven patients experienced 16 pregnancies complicated by spontaneous abortion or stillbirth (31%), dysmaturity (39%) and postpartum syncope, hemorrhage or sustained symptomatic deterioration. Women with the Eisenmenger syndrome should avoid pregnancy or accept early termination (2,10). Experience in our Center, in addition to this report, supports a consensus that maternal risk is reduced by planned, induced, controlled vaginal delivery in a specialized high risk obstetrical facility, and that risk is increased by cesarean delivery, which should be reserved for obstetrical indications (2,10).
In group A, infective endocarditis did not occur despite moderate to severe tricuspid and/or pulmonary regurgitation. Inherently normal pulmonary or tricuspid valves rendered incompetent by pulmonary hypertension appear to incur little or no risk, and low velocity right to left shunts through nonrestrictive VSDs are not risk substrates (11). In group B, there were no cases of infective endocarditis despite abnormal truncal valves. In group C, infective endocarditis occurred in one patient on the right atrioventricular valve, and in another patient on the left atrioventricular valve. Both valves were inherently malformed and incompetent before infection.
Morbidity reflected the risk of PVD addition to the risk of abnormal hemostasis in cyanotic patients. Risks of “minor” surgery were underscored in two patients in whom serious hemorrhage was induced by dental procedures (12,13). An experienced cardiac anesthesiologist is as important as a skilled surgeon in risk reduction (12,13). Additional precautions include preoperative phlebotomy to improve hemostasis by reducing hematocrit levels to just below 65% (5), avoidance of percutaneous flotation catheters, insertion of an air/particle filter into the distal end of intravenous lines, intraoperative and postoperative monitoring of systemic vascular resistance and venoarterial shunting utilizing a fingertip oximeter and attention to postoperative postural hypotension.
Multisystem systemic disorders
A major cause of hematologic morbidity was iron deficiency induced by inappropriate phlebotomy. Erythrocytosis is a desirable adaptive response to hypoxemia in cyanotic congenital heart disease (5)and does not incur risk of stroke due to cerebral arterial thrombosis (14). Phlebotomy should be reserved for temporary relief of major hyperviscosity symptoms (5,14), or to improve perioperative hemostasis (see preceding text).
Bleeding tendencies, apart from pulmonary hemorrhage, were mucocutaneous and were represented by epistaxis, gingival bleeding especially during dental procedures and cutaneous bruising both spontaneous and in response to minor trauma (15). Aspirin and nonsteroidal anti-inflammatory agents reinforce hemostatic defect(s) and are best avoided (5,15).
Oxygen inhalation was not used because inhaled oxygen leaves the saturation of shunted blood unaffected, and because the drying effect of nonhumidified oxygen predisposes to epistaxis (5).
Cholecystitis caused by calcium bilirubinate gallstones was the commonest reason for major noncardiac surgery. Bilirubin is formed from the breakdown of heme, a process that is excessive in the presence of erythrocytosis, and that results in substantial increases in unconjugated bilirubin, which is virtually water insoluble at physiologic pH (12). Laparoscopic cholecystectomy should be used cautiously because of potential difficulty in the control of bleeding (13).
Hyperuricemia occurred in over 50% of our patients, with morbidity expressed chiefly as gouty arthritis, the average incidence of which was 20%. Elevated plasma uric acid levels, which are secondary to reduced renal clearance (16), seldom adversely affect renal function (Table 3), so asymptomatic hyperuricemia was not routinely treated (5). Normal renal function was not altered by the presence of gout, the prevalence of which was equal in men and women. Gout had its onset between the third and fifth decades, and was not necessarily related to uric acid levels.
Proteinuria occurred in 44% to 74% of patients, but serum albumin levels were consistently normal (Table 3). The glomerulus is porous to albumin which is nevertheless retained because glomerular capillary walls are negatively charged, whereas the protein molecule is cationic (17). Erythrocytosis increases perfusion pressure that overcomes the cationic effect, so albumin leaves the glomerular circulation (17). Because albuminuria with hypoalbuminemia is associated with hypercoagulability (18), documentation of normal serum albumin levels in our patients was an important observation.
Despite the frequency of albuminuria and hyperuricemia, renal function as judged by blood urea nitrogen and serum creatinine levels was normal (Table 3). Exceptions were two group A patients, one of whom experienced chronic renal failure.
Morbidity in the pulmonary circulation in addition to PVD was expressed as pulmonary hemorrhage (Fig. 2), and as thromboses, calcification and aneurysms of hypertensive proximal pulmonary arteries (Fig. 3). Hemoptysis, which recurred in 21% to 57% of patients, is by definition external, and did not reliably reflect the presence or extent of intrapulmonary hemorrhage. When hemoptysis was accompanied by pulmonary infiltrates on chest X ray, a CT scan established the degree of intrapulmonary hemorrhage (Fig. 2a), which can be extensive and occasionally fatal (see below) (Fig. 2b). We recently identified an acquired von Willebrand abnormality in the Eisenmenger syndrome (15), but the efficacy and safety of desmopressin (2,15)and human factor VIII (15), which have been considered potentially beneficial in the control of hemorrhage, have not been tested.
The incidence and extent of large in situ thromboses within dilated hypertensive proximal pulmonary arteries (Fig. 3a and b)justify CT scans as part of the diagnostic assessment of adults with the Eisenmenger syndrome. However, there are no therapeutic modalities that satisfactorily address these in situ thromboses. In our experience, intrapulmonary thrombolytic agents are not effective, and anticoagulants are more likely to cause hemorrhage than prevent progression (5). When proximal thrombotic material embolized into distal pulmonary arteries causing hemorrhagic pulmonary infarction, we cautiously used short-term low dose coumadin in an attempt to stabilize the proximal thromboses.
Morbidity in the pulmonary circulation was also represented by mural calcification and aneurysmal dilation of proximal pulmonary arteries, which occasionally reached extraordinary proportions (Fig. 3a and 4a), causing symptomatic bronchial compression or sudden death due to rupture (Fig. 4b). Proximal pulmonary arterial thromboses (Fig. 3), heavy mural calcification (Fig. 3a)and aneurysmal dilation (Fig. 3a)warn against single-lung transplantation.
Long bone morbidity, signaled by symptomatic hypertrophic osteoarthropathy, occurred in 8% to 12% of our patients. Asymptomatic hypertrophic osteoarthropathy might have been detected by systematic palpation of long bones, by X rays or more precisely by technetium scans (5).
Morbidity caused by cerebral disorders was represented by three remote brain abscesses in two patients. Neither had seizure disorders. Cerebral arterial thrombotic stroke due to erythrocytosis per se did not occur irrespective of hematocrit level or iron stores.
Pregnancy incurred considerable morbidity. Early termination, or planned, controlled, induced vaginal delivery are important in risk reduction (10). Noncardiac surgical morbidity was reduced by a cardiac anesthesiologist and by meticulous medical management (12,13). Strict criteria for phlebotomy avoided deleterious effects of induced iron deficiency (5,14). Aspirin or other anti-inflammatory agents reinforce intrinsic hemostatic defects (5,15). Avoidance of nonhumidified oxygen inhalation reduced the risk of epistaxes. Detection of intrapulmonary hemorrhage by CT scan sets the stage for fresh frozen plasma, cryoprecipitate or platelet transfusion. Anticoagulants are of no therapeutic benefit for proximal pulmonary arterial thromboses, but instead increase morbidity from bleeding.
Mean collective age of survival in early reports of the Eisenmenger syndrome was in the late teens or twenties (1,2,19–26). Longevity in our patients was approximately one to two decades greater. Mean ages of our living patients were: group A 37 ± 8.1, group B 31 ± 6.3 and group C 33 ± 6 years. Mean ages at time of death were: group A 45 ± 5.0, group B 41.5 ± 5.1 and group C 31 ± 6.6 years. The 3- to 5-year actuarial survival rates after initial visit were significantly different in groups A, B and C (Fig. 5). Longevity was shorter in group B (truncus) than in group A (VSD), probably because of acquired regurgitation or stenosis of the biventricular truncal valve. Longevity in group C was shorter than in either group A or B, probably because of inherently poorer function of the UVH (6,7), especially those of right ventricular morphology (8)(see beginning of paragraph). Infants with unoperated TA were less likely to survive than were infants with isolated nonrestrictive VSD (3). If adult survival was achieved, differences in longevity narrowed but were not eliminated.
Twenty-seven of 77 patients (35%) died. Seventeen of the 27 deaths (63%) were sudden and occurred in patients who were functional class II or III and were working immediately before death. Tachyarrhythmic sudden death was not documented, but was not excluded when the cause of sudden death was unestablished. Six patients experienced nonsustained monomorphic ventricular tachycardia, and two experienced sustained nontachycardiac monomorphic ectopic ventricular rhythms (cardioverted), but none of these eight patients died suddenly. Instead, sudden death was due to a variety of causes, namely, massive intrapulmonary hemorrhage (Fig. 2b), rupture of an aneurysmal pulmonary trunk (Fig. 4b), dissection of the ascending aorta, rupture of a bronchial artery, vasospastic cerebral infarction (migraine) or unestablished.
A significant increase in life span in our patients compared with earlier reports may have been due to management of atrial and ventricular tachyarrhythmias and management of pregnancy and noncardiac surgery. Improved longevity exposed patients to the risk of nontachyarrhythmic sudden death, the relatively high incidence of which clouds the planning of lung or heart/lung transplantation.
In the Eisenmenger syndrome, represented by nonrestrictive VSD, TA or UVH, morbidity and mortality patterns differ despite a systemic ceiling on ventricular systolic pressure. Medical management of coexisting cardiac disease, multisystem systemic disorders, noncardiac surgery and pregnancy has reduced morbidity. Life span in group B patients was shorter than in group A, and life span in group C was shorter than in either group A or B. Overall longevity was approximately one to two decades longer than in earlier reports, but longer life spans exposed patients to proximal pulmonary arterial aneurysms, thromboses and calcification; to truncal valve regurgitation and stenosis; to semilunar and atrioventricular valve regurgitation, and to a substantial risk of nontachyarrhythmic sudden death.
We thank Alan Garfinkel, PhD for his review of the statistical analyses.
- computerized tomography
- pulmonary vascular disease
- truncus arteriosus
- univentricular heart
- ventricular septal defect
- Received June 8, 1998.
- Revision received February 17, 1999.
- Accepted March 15, 1999.
- American College of Cardiology
- ↵Wood P. The Eisenmenger syndrome: or pulmonary hypertension with reversed central shunt. Br Med J 1958;2:701–9, 755–62.
- Perloff J.K
- ↵Child JS. Transthoracic and transesophageal echocardiographic imaging: anatomic and hemodynamic assessment. In: Perloff JK, Child JS. Congenital Heart Disease in Adults. 2nd ed. Philadelphia: WB Saunders, 1998:91–128.
- ↵Perloff JK, Rosove MH, Sietsema KE, Territo MC. Cyanotic congenital heart disease: a multisystem systemic disorder. In: Perloff JK, Child JS. Congenital Heart Disease in Adults. 2nd ed. Philadelphia: WB Saunders, 1998:199–226.
- Fogel M.A,
- Weinberg P.M,
- Fellows K.E,
- Hoffman E.A
- Fogel M.A,
- Weinberg P.M,
- Gupta K.B,
- et al.
- Sano T,
- Ogawa M,
- Yabucchi H,
- et al.
- ↵Perloff JK, Koos B: Pregnancy and congenital heart disease. In: Perloff JK, Child JS. Congenital Heart Disease in Adults. 2nd ed. Philadelphia: WB Saunders, 1998:144–64.
- ↵Perloff JK, Sangwan S. Noncardiac surgery. In: Perloff JK, Child JS. Congenital Heart Disease in Adults. 2nd ed. Philadelphia: WB Saunders, 1998:291–9.
- Ammash N.M,
- Connolly H.M,
- Abel M.D,
- Warnes C.A
- Perloff J.K,
- Marelli A.J,
- Miner P.D
- Territo M.C,
- Perloff J.K,
- Rosove M.H,
- et al.
- Ross E.A,
- Perloff J.K,
- Danovitch G.M,
- et al.
- Clarkson P.M,
- Frye R.L,
- DuShane J.W,
- et al.
- Marcelletti C.M,
- McGoon C.M,
- Mair D.D
- Brammell H.L,
- Vogel J.H.K,
- Pryor R,
- Blount S.G Jr.