Author + information
- Received June 17, 1999
- Revision received April 3, 2000
- Accepted April 11, 2000
- Published online September 1, 2000.
- Thomas Gilljam, MD, PhD∗,†,
- Brian W McCrindle, MD, MPH, FACC∗,†,
- Jeffrey F Smallhorn, MBBS, FACC∗,†,
- William G Williams, MD, FRCS(C)∗,† and
- Robert M Freedom, MD, FACC∗,†,* ()
- ↵*Reprint requests and correspondence: Dr. Robert M. Freedom, the Hospital For Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada
We determined long-term outcomes in a large cohort with left atrial isomerism (LAI).
Left atrial isomerism is associated with a complex spectrum of cardiac and noncardiac anomalies that may impact on outcomes.
The records of all patients with LAI, born between 1970 and 1998, and treated at one center were reviewed. Kaplan-Meier survival was estimated, and independent factors associated with time-related death were identified.
There were 163 patients (63% women), and extracardiac anomalies were noted in 36%, including biliary atresia in 10%. Cardiac defects included interrupted inferior caval vein in 92%, anomalous pulmonary veins in 56%, atrioventricular septal defect in 49%, pulmonary atresia or stenosis in 28% and aortic coarctation in 16%, with congenital atrioventricular block in 7%. Of 22 patients with a normal heart, 18% died of extracardiac anomalies. Of 71 patients with hearts suitable for biventricular repair, 62 (87%) had surgery, with survival of 80% at one year, 71% at five years, 66% at 10 years and 63% after 15 years. Of 70 patients with unbalanced cardiac defects suitable for single-ventricle palliation, 47 (67%) had surgery, with survival of 73% at one year, 61% at five years, 53% at 10 years and 48% at 15 years (p < 0.001). Independent factors associated with time-related death included congenital atrioventricular block, aortic coarctation, single ventricle, biliary atresia and other gastrointestinal malformations.
Both cardiac and noncardiac anomalies contribute to a high mortality with LAI. Cardiac transplantation may need to be a considered a primary option for selected high-risk patients.
Left atrial isomerism (LAI), formerly designated “polysplenia syndrome” by some, has been recognized as a clinical entity for 30 years, and the spectrum of associated cardiac and noncardiac anomalies has been thoroughly delineated (1–4). In the past the often complex systemic and pulmonary venous connections and intracardiac anatomy made long-term palliation or repair difficult, and most patients died during the first year of life (2,4). More recently, methods for surgically addressing these complex anomalies have been developed (5–9), with subsequent reports on successful palliation or repair in small to medium-sized case series (6–8,10).
Since data on outcome in a larger population with LAI are lacking, we have reviewed our recent 28-year experience, comprising 163 patients. We included surgical as well as autopsy cases and patients with normal (N) hearts. We also sought to determine factors associated with mortality in this diverse group of patients.
We searched the databases of the Division of Cardiology, Department of Pediatrics, the Department of Cardiovascular Surgery and the Department of Pathology at the Toronto Hospital For Sick Children for cases of LAI born between January 1970 and June 1998. The search diagnoses are listed in Table 1. Patients were considered for inclusion in the study after reviewing the charts for these diagnostic features. The diagnoses were made using a variety of investigations. Descriptions of anatomy, cardiac function and dysrhythmias were abstracted from records of surgery, autopsy, cardiac catheterization with angiography, chest X-ray, echocardiography and, more recently, fetal echocardiography and magnetic resonance imaging. Additional information on abdominal anatomy was obtained by ultrasound and scintigraphy (“spleen scan”). Three potential cases were excluded due to a lack of sufficient data.
Left atrial isomerism was confirmed if there was abdominal situs ambiguous, with or without polysplenia, with bilateral hyparterial bronchi or bilobed lungs or confirmed bilateral left atrial appendage morphology. A probable diagnosis was considered if there was an interrupted inferior caval vein with at least two other features of LAI, for example, left atrial rhythm (defined as a P-wave vector of 180 to 360°), biliary atresia or congenital heart disease within the spectrum of atrioventricular (AV) septal defects.
Demographic, anatomic, procedural and outcomes data were included in the analysis. Survival beyond last hospital visit or contact was not assumed. Results are described as frequencies, means with standard deviation or medians with range where appropriate. Where there are missing data, the number of nonmissing values is given. Categorical variables were analyzed using the Fisher exact test. Survival was estimated according to the method of Kaplan and Meier, and differences between groups were analyzed using log-rank and Wilcoxon tests. Factors associated with a time-related event (death) were analyzed using Cox’s proportional hazard regression modeling using default settings. Variables that were related to mortality with p < 0.10 in univariate analysis were entered into a multivariable model with stepwise forward selection, with entry criteria of p < 0.05. Statview 5.0 software (SAS Institute, Cary, North Carolina) was used for all analyses.
There were 163 patients, 14 of whom were diagnosed at autopsy. In 20 patients (12%) the diagnosis was made or suspected at fetal echocardiography. (There were 14 additional fetal cases who subsequently experienced termination of pregnancy or fetal death.) The population constituted 0.4% of all new patients with congenital heart disease diagnosed at our institution during the study period. There were 102 women (63%) and 61 men. The mean gestational age was 38 ± 2 weeks (n = 123), and the mean birth weight was 3.1 ± 0.7 kg (n = 123). In 18 patients (15%) the birth weight was less than 2.5 kg. In 10 patients (6%), it was recorded that the mothers had diabetes mellitus.
The median age at presentation (n = 161) was two days (range < 1 day to 16 years). At presentation, 120 patients (74%) had cardiac symptoms: 75 patients (46%) with congestive heart failure and 55 patients (34%) with cyanosis. In 26 patients (16%), incidental findings, for example, a murmur or dextrocardia, led to the diagnosis. In six patients with biliary atresia (4%), there were symptoms of hepatic failure at presentation. In 11 additional patients (7%) the presentation was mainly related to noncardiac features: myelomeningocele in two patients, sepsis in one, symptoms of gastrointestinal malrotation in two, respiratory in four, failure to thrive in one and omphalocele in one patient.
Cardiac anatomy is summarized in Table 2. There was levocardia in 96 patients (62%, n = 156), dextrocardia in 46 (29%) and mesocardia in 14 patients (9%).
The pulmonary venous return (n = 155) was partially anomalous to bilateral atria in 77 patients (50%); one patient had a Scimitar syndrome. There was total anomalous pulmonary venous drainage in 13 patients (8%): to the right-sided atrium in 12 patients and to the coronary sinus in one patient. Mild pulmonary venous obstruction was present in three patients. In 64 patients (41%) no pulmonary venous abnormalities could be confirmed.
The inferior caval vein (n = 162) was interrupted in 149 patients (92%), with an azygous vein continuation to a right-sided superior cava in 97 patients (65%), to a left-sided superior cava in 51 (31%) and connection into the coronary sinus in one patient. The superior caval vein (n = 162) was right-sided in 70 patients (43%), left-sided in 24 (15%) and bilateral in 68 patients (42%). In four patients, the left superior caval vein drained into the coronary sinus. The hepatic venous return (n = 161) was to the inferior caval vein in seven patients (4%) and directly to the atrium in 154 patients (96%).
The atrial septum was intact in 27 patients (17%), and it was virtually absent (a common atrium) in 82 patients (50%). One patient had a sinus venous defect, and one had a divided left-sided atrium. In the remainder (52 patients, or 32%), there was either a smaller defect of the secundum type or a patent foramen ovale. The coronary sinus (n = 56) was absent or unroofed in 20 patients (36%) and normal in the 36 patients (64%).
Mitral valve anomalies were found in 10 patients (6%): an absent connection in four patients, a parachute mitral valve in two, a dysplastic valve in three and a supravalvar ring in one patient (Shone’s complex). Tricuspid valve anomalies were present in four patients (2%): tricuspid atresia in two and a dysplastic tricuspid valve in two patients. An AV septal defect was present in 80 patients (49%), seven of whom had an intact ventricular septum (a primum atrial septal defect).
A ventricular septal defect was diagnosed in 111 patients (68%). In 70 of them (63%), the defect was unbalanced with one hypoplastic ventricle, 52 of whom had a double outlet right ventricle. Furthermore, 27 patients had ventriculoarterial discordance.
Symptomatic pulmonary outflow obstruction was present in 22 patients (13%) and pulmonary atresia in an additional 23 patients (14%). Pulmonary artery anomalies were present in 34 patients (21%): hypoplastic pulmonary arteries in 25, nonconfluent pulmonary arteries in 4, a truncus arteriosus in 1, a pulmonary sling in 1, a right pulmonary artery from aorta (hemitruncus) in 1 and absence of one lung in 2 patients.
Coarctation or hypoplasia of the aortic arch was present in 27 patients (17%), 16 of whom had considerable obstruction to aortic outflow and 13 of whom had a severely hypoplastic subaortic ventricle. One of them also had a complex double aortic arch with interruption.
The abdominal situs was inversus in 56 (40%, n = 139) and indeterminate with a midline liver in 50 patients (36%). In 33 patients (24%), sufficient information was not recorded to accurately define the abdominal situs. There was confirmed polysplenia in 57 patients (59%, n = 96), with the number of spleens or nodules ranging from 3 to 15 in 17 autopsy cases. In 26 patients (27%), the spleen was diagnosed as normal, and in 13 patients (14%) it was either hypoplastic or absent, with findings in only 2 of these 39 cases being confirmed by a postmortem investigation. In 51 of 53 patients (96%) where pulmonary anatomy was investigated, there were bilateral hyparterial bronchi or bilobed lungs.
Extracardiac anomalies and conditions were diagnosed in 58 patients (36%) (Table 3). A portojejunostomy was performed in 11 of 16 patients with biliary atresia (69%), and 4 of them proceeded to have a liver transplantation. A renal transplantation was later performed in a patient with renal glomerulosclerosis. Other noncardiac operations were: tracheoesopageal fistula repair in one patient, relief of volvulus due to malrotation of the gut in four patients, in three patients covering of a myelomeningocele, in two patients a ventriculoperitoneal shunt for hydrocephalus, in four patients repair of cleft lip/palate, in one patient esophageal atresia repair, in two patients covering of an omphalocele and in one patient repair of a diaphragmatic hernia.
Electrocardiographic findings and dysrhythmias
At presentation or during follow-up, a left atrial rhythm was evident in 96 patients (63%, n = 153). A wandering atrial pacemaker, defined as a change in the P-wave vector of 90° or more between electrocardiographic tracings, was found in 23 patients (14%).
Dysrhythmias were noted in 43 patients (26%). Loss of atrial rhythm with bradycardia developed in 26 patients (16%), three of whom later acquired atrial flutter and one of whom also developed ventricular tachycardia. Congenital complete AV block was present in 11 patients (7%). Postoperative AV block developed in four patients. Two patients had episodes of supraventricular tachycardia. In total 20 patients (12%) received a pacemaker at a median age of 6 months (range < 1 day to 16 years), 9 for loss of atrial rhythm and 11 for complete AV block.
For the purpose of this study, we divided the 163 patients into three groups based on their cardiac anatomy: those with normal (N) hearts (defined as with or without an interrupted inferior caval vein or insignificant atrial or ventricular septal defects and valvular anomalies), those patients with balanced defects who were suitable for biventricular (BV) repair and those with unbalanced defects only suitable for palliation or single-ventricle (SV) “repair.” There were 22 patients with N hearts (13%), 71 with BV (44%) and 70 patients (43%) with SV hearts (Fig. 1). The proportion of patients with N hearts increased somewhat during the study period.
No cardiac surgery
In five patients with BV heart and 22 patients with SV heart, there was an early death, either before any interventions could be performed or they were deemed inoperable due to the complexity of their cardiac defect or associated extracardiac anomalies. Only three of these patients are known to be alive at 0.7, 1.4 and 5.4 years, respectively, and all with considerable symptoms. In seven of the patients who died (26%), extracardiac factors primarily contributed: prematurity or low birth weight, sepsis, biliary atresia and heart perforation during cardiac catheterization. In five patients with congenital AV block (four from the BV and one from the SV group), an urgent pacemaker implantation was performed during the first week of life. Of these patients, four died before other interventions were attempted due to heart failure, pacemaker failure and hydrops foetalis, respectively. Another patient was subsequently denied surgery for an AV septal defect due to the presence of severe pulmonary vascular disease. Over the study period, there was no apparent decrease in the proportion of patients who did not have any treatment.
In the group with an N heart, four patients (18%) died: three in conjunction with liver transplantation for biliary atresia at 5 to 13 months of age, and one patient died at one month of age due to complications from arthrogryposis. The remaining 109 patients entered into a protocol for BV repair or SV palliation (Fig. 2).
Of 47 patients who entered the SV protocol, 43 received a primary palliation. An aortopulmonary shunt was carried out in 31 patients, together with atrial septectomy in 4, hemitruncus repair in 1 and Norwood operation in 3 patients with a mortality before 30 days after surgery of 23% (7 patients). Pulmonary artery banding was performed in 10 patients, together with coarctation repair in 3 and aortic stenosis repair in 2 patients, with an operative mortality of 30% (3 patients). Two patients who had creation of an atrial septum defect, one of whom, who also had pulmonary outflow reconstruction, subsequently died.
In these first palliations, causes of death included heart failure (one patient), systemic arterial shunt complications (three patients), sepsis (two patients), complications from tracheoesophageal fistula (one patient), ventricular tachycardia (one patient), necrotizing enterocolitis (one patient), complications from prematurity (one patient) and apnea due to Arnold-Chiari malformation (one patient).
A bidirectional cavopulmonary shunt was performed in 20 patients (bilateral in 6 patients) and a classic atrium to pulmonary artery connection in 4 patients (mortality 8% or 2 patients). Subsequently, 10 patients had the hepatic veins diverted to the pulmonary artery by an extracardiac conduit, with a mortality of 50% (five patients). The causes of death included multiorgan or heart failure (three patients), postoperative bleeding (two patients), junctional ectopic tachycardia and severe shunting through arteriovenous fistulas (one patient each). There was one late death after placement of a bidirectional cavopulmonary shunt due to poor cardiac function in a patient with severe neurological impairment and a chromosomal abnormality.
Of the remaining 12 survivors of first palliation (who did not have further surgery), four died during the first postoperative months due to heart failure, sepsis and pulmonary venous occlusive disease, respectively. The other eight patients are alive; a heart transplantation has been performed in two of these patients. The remaining six patients have not been considered as candidates for further surgery due to unfavorable cardiac findings or pulmonary vascular disease.
Thus, among the 47 patients operated on, 24 have survived. Among the deaths, 14 (61%) were directly attributed to operations. Overall, for the 70 patients in the SV group, 22 had no treatment (20 died, 2 alive but severely limited), 1 had AV block and a pacemaker (died) and 47 had cardiac surgery (23 died and 2 had heart transplantation), giving a total of 44 deaths (63%).
Of 62 patients entering the BV track, 18 had a primary palliation involving a pulmonary artery band in 9, an aortopulmonary shunt in 6 or a ligation of the ductus arteriosus in 3, with an operative mortality of 6% (1 patient). During the first months of life, three of the survivors died before additional surgery could be performed (see below). Another patient has been denied further surgery due to severe pulmonary vascular disease.
The remaining 57 patients had cardiac repair, a two-stage repair in four patients, involving initial aortic coarctation repair in two patients and pulmonary vein baffling or pulmonary valvotomy in one patient, respectively. Generally, the other operations involved closure of existing atrial and ventricular septal defects and repair of AV septal defects and coarctation of the aorta. Additional procedures performed included baffling of partial anomalous pulmonary veins (24 patients) or total anomalous pulmonary veins (3 patients), aortic outflow repair (8 patients), pulmonary outflow reconstruction (9 patients), redirection of left superior vena cava (5 patients) and AV valve replacement, arterial switch, repair of pulmonary sling or mitral stenosis (1 patient each).
Of the survivors 11 patients had reoperations for mitral or aortic valve regurgitation, residual coarctation or ventricular septal defect for pulmonary venous obstruction with pulmonary edema or pulmonary stenosis.
There were 12 perioperative deaths (19%) and five late deaths in the group. The causes of death included congestive heart failure (seven patients), complications related to prematurity (one patient), bleeding (one patient), undiagnosed total anomalous pulmonary venous return (one patient) and hepatic failure in two patients with biliary atresia. The five late deaths all occurred during the first months of life, in three patients attributed to the biliary atresia, in one from complications of myelomeningocele and in one patient from complications of pulmonary hypertension.
Factors associated with mortality
Overall survival was 56% (91 patients), 82% in the N group (18 patients), 66% in the BV group (47 patients) and 37% in the SV group (26 patients). Kaplan-Meier estimates for the entire cohort was 80% at 1 month, 68% at 1 year, 59% at 5 years, 55% at 10 years and 51% at 15 years (Fig. 3). Figure 4 depicts survival estimates for patients with N, BV and SV hearts. The difference in survival between groups was significant (p < 0.001, log-rank and Wilcoxon tests). There was somewhat improved survival during the second half of study period (1985 to 1998) than during the first (1970 to 1984), but the difference was not significant (p = 0.16, patients with N hearts excluded). Only 1 of 11 (9%) of the patients with congenital AV block is known to be alive. Exclusion of the patients who were refused treatment improved the survival estimates by less than 5%.
At follow-up including echocardiography, 61 of the survivors (67%) had qualitatively good cardiac function and were on no anticongestive medications, 20 had somewhat decreased cardiac function or required anticongestive medication, and 10 patients were clearly incapacitated from a cardiac standpoint. Furthermore, eight patients had a pacemaker, five had neurologic abnormalities ranging from hyperactivity disorder to severe neurologic deficits, two had cardiac transplantation, three had respiratory problems, two had hypothyroidism and all survivors with biliary atresia (5/16 or 31%) had some morbidity due to medication or decreased liver function.
In the univariate analysis (Table 4)gestational age, birth weight, age at presentation, SV heart, AV valve and pulmonary vein anomalies, coarctation of the aorta, aortic stenosis, ventriculoarterial discordance, congenital AV block, biliary atresia, gastrointestinal abnormalities other than biliary atresia and age at first intervention were associated with mortality with p < 0.10. The age at presentation as well as the age at the first intervention were excluded from further analyses since these factors were significantly related to anatomic factors. In the multivariate analysis model, with entry criterion of p < 0.05, only the variables SV heart, coarctation of the aorta and birth weight were retained (Table 5). Since there was important missing data for birth weight, the model building was repeated without this variable; in this model biliary atresia and other gastrointestinal abnormalities and congenital AV block were also retained.
This is, to the best of our knowledge, the first paper depicting the outcomes in a large single-institution cohort with LAI.
The cohort constituted 0.4% of all children with heart disease presenting to our institution, which is somewhat lower than incidence data reported by others (1 to 2%) (11,12). Of 34 fetal cases diagnosed since 1988, 14 (41%) underwent termination of pregnancy or died in utero. This confirms previous data (13,14) on a relatively poor outcome of LAI detected during fetal life and suggests that the fetal prevalence is considerably higher than the postnatal.
We confirmed previous reports of a female preponderance (15) and a similar spectrum and relative frequency of associated cardiac defects (2,4,7,16) and extracardiac anomalies and conditions (2,4,17). The relative frequency of some asymptomatic extracardiac malformations, for example, polysplenia and malrotation of the gut, is likely to be underestimated since many patients were not properly investigated for these anomalies. An association between atrial isomerism and maternal diabetes has been suggested by Slavotinek (18). In this study at least 6% of the mothers were reported to have diabetes.
During the study period, 20% of patients died shortly after birth or were not considered eligible for surgery due to the presence of complex cardiac lesions, associated anomalies or other conditions, such as prematurity. Comparable data are not available but in a smaller reported cohort of 27 patients from 1970 to 1984 (15) 49% did not have surgery, and most died within the first year of life, confirming earlier autopsy data (2,4). It has been reported that at least 10% of patients with LAI have a normal heart (2), confirmed by our estimate of 13%. These patients may not be diagnosed until they are of adult age (19) and have a normal life span (2), unless they have associated extracardiac conditions (17), as noted by the 18% mortality in this group in our study.
Methods for repairing or palliating complex heart defects associated with LAI have gradually developed during the study period (5–9,20), and, recently, cardiac transplantation has also been performed in patients either unsuitable for surgery or after unsuccessful surgery, as highlighted by two successful cardiac transplantations in this series.
In the BV repair group, the surgical mortality was 19%. When deaths that were unrelated to surgery were excluded, the mortality was 15%, compared with a mortality of 7% reported in another recent series (6) of 15 patients who underwent BV repair.
The total surgical mortality in the SV group was 38%. When deaths that were unrelated to surgery were excluded, the mortality was 30%. The relative high mortality with the initial palliation (26%) can be explained by the fact that 32% of these operations involved complex procedures. Comparable data are not reported, but mortality in simple aortopulmonary shunt placement in LAI are reported to be low (15,20). The mortality with cavopulmonary shunt placement in patients with interrupted inferior caval vein with azygous connection, the so-called Kawashima operation (5), in our study was 8% and comparable to other reports of 0% to 33% (5,21,22). More recently there have been reports of the development of pulmonary arteriovenous fistulas, due to the exclusion of hepatic venous blood from the pulmonary circulation, after the Kawashima operation (23,24). Therefore, we, and others (23,24), have chosen to complete the Fontan circulation in these patients, for example, by an extracardiac conduit (9), and there have been reports of subsequent resolution of intrapulmonary shunts (23,24). The mortality with Fontan operations in LAI has been reported to range from 0% to 27% (10,25), with an improvement over time noted in the study of Cetta (10), compared with the mortality of 29% up to a 28-year-period in our study, albeit with a mortality of 50% for the extracardiac Fontan operation to connect the hepatic veins into the circuit.
Factors associated with mortality
In multivariate analysis, three cardiac factors (AV block, SV and coarctation of the aorta) and three extracardiac (biliary atresia, other gastrointestinal malformations and low birth weight) were independently associated with increased mortality. The analysis including birth weight should be interpreted with caution due to the important lack of data, which may have introduced a bias. In contrast with right atrial isomerism (26), abnormal pulmonary venous connections were not undisputedly associated with an increased risk.
The outcome in LAI with fetal AV block is reported to be poor, with a majority of the patients dying before or shortly after birth (14,27). The late outcome in the early survivors is reported to be poor, in spite of pacing (28). Indeed, in our study 4 of 5 patients (80%) who had early pacemaker therapy died before further surgery could be performed, and only one of the total cohort of 11 patients (9%) survived. Nevertheless, congenital AV block was not retained in the model that included birth weight; this might be explained by the fact that the median birth weight in patients with AV block was considerably lower than it was in the overall study population (2.7 vs. 3.1 kg).
Survival for patients with biliary atresia was low (31%) in spite of 69% having a portojejunostomy, some of whom later had a liver transplantation. In contrast, Vasquez (29) reported a 72% survival in 11 patients with biliary atresia and polysplenia, albeit in a selected group where only two patients had cardiac defects.
It is not entirely clear why gastrointestinal abnormalities other than biliary atresia (for example, omphalocele, esophageal atresia and malrotation of the gut) were associated with an increased risk since only one of the deaths was directly attributed to such anomalies, and only in two patients was treatment deferred due to such anomalies.
It is less surprising that the presence of an SV was associated with increased mortality. We have, in fact, shown that the survival for this subgroup of patients is on a par with the five-year survival data of 35% in cases with right atrial isomerism in a recent study from our institution (26). This is in contrast with previous suggestions, that the prognosis is better for patients with LAI than right atrial isomerism (4). Comparably better survival in the BV group may be attributed to a lower operative mortality (6). Coarctation of the aorta was also associated with less optimal outcome, which may be expected since several of these patients had associated aortic outflow obstruction and hypoplastic left ventricle.
Our study highlights the diversity in presentation, anatomy and clinical outcomes of LAI. Associated cardiac defects, such as SV and aortic coarctation, congenital AV block and extracardiac anomalies, such as biliary atresia, were significantly associated with increased mortality. Recent advances in cardiac surgery and liver transplantation may offer improved prospects for survival in current patients with complex heart disease and biliary atresia (29). In patients with congenital AV block and complex heart disease, the outlook may still be unfavorable, and in these patients cardiac transplantation may be offered as an alternative option.
☆ Supported by the Swedish Institute, the Swedish Medical Society and the Gothenburg Medical Society.
- left atrial isomerism
- single ventricle
- Received June 17, 1999.
- Revision received April 3, 2000.
- Accepted April 11, 2000.
- American College of Cardiology
- Moller J.H,
- Nakib A,
- Anderson R.C,
- Edwards J.E
- van Mierop L.H.S,
- Gessner I.H,
- Schiebler G.L
- Rose V,
- Izukawa T,
- Moes C.A.F
- Kirklin JW, Barratt-Boyes BG. Atrial isomerism. In: Cardiac Surgery. 2nd ed. Edinburgh: Churchill-Livingstone, 1993:1585–91.
- Rosenkranz E.R,
- Murphy D.J Jr.
- Cetta F,
- Feldt R.H,
- O’Leary P.W,
- et al.
- Sharma S,
- Devine W,
- Anderson R.H,
- Zuberbuhler J.R
- Correa-Villasenor A,
- McCarter R,
- Downing J,
- Ferencz C
- ↵Phoon CK, Villegas MD, Ursell PC, Silverman NH. Left atrial isomerism detected during fetal life. Am J Cardiol 1996:1083–8.
- Macahdo M.V.L,
- Tynan M.J,
- Curry P.V.L,
- Allan L.D
- Culbertson C.B,
- George B.L,
- Day R.W,
- Laks H,
- Williams R.G
- Hashmi A,
- Abu-Sulaiman R,
- McCrindle B.W,
- Smallhorn J.F,
- Williams W.G,
- Freedom R.M
- Schmidt K.G,
- Ulmer H.E,
- Silverman N.H,
- Kleinman C.S,
- Copler J.A