Author + information
- Received July 22, 1999
- Revision received October 5, 1999
- Accepted January 12, 2000
- Published online May 1, 2000.
- Teri Millane, MD, MRCP∗,
- Elizabeth J Bernard, FRACP†,
- Edgar Jaeggi, MD∗,
- Robert B Howman-Giles, MD, FRACP, DDU†,
- Roger F Uren, FRACP, DDU†,
- Timothy B Cartmill, FRACS‡,
- Richard E Hawker, FRACP∗ and
- David S Celermajer, PhD, FRACP∗,§
This study was conducted to assess whether myocardial ischemia and/or infarction are involved in the pathogenesis of late right ventricular dysfunction in adult survivors of atrial baffle repair for transposition of the great arteries in infancy.
The medium-term success of intraatrial baffle repair for transposition of the great arteries is good, with many patients surviving into adult life, but prognosis can be limited by progressive right ventricular dysfunction. We hypothesized that ongoing myocardial ischemia and/or infarction are important factors in the pathogenesis of this complication. Radionuclide techniques offer an opportunity to study both myocardial perfusion and concomitant ventricular wall motion.
Dipyridamole sestamibi single-photon emission computed tomography followed by rest sestamibi single-photon emission computed tomography was used to assess right ventricular myocardial perfusion, wall motion, wall thickening and ejection fraction in 22 adolescents/young adults who had undergone atrial baffle repair for simple transposition of the great arteries at median 6.7 (range 0.5 to 54) months of age. The patients were aged 10 to 25 (median 15.5) years; 19 in New York Heart Association class I, 2 in class II and 1 in class III. All were in a regular cardiac rhythm during the studies. The right ventricular tomographic images were examined in three parallel and two orthogonal planes, analyzed in 12 segments.
Perfusion defects were evident in all patients in at least one segment, in either the rest or stress images. Twelve patients (55%) demonstrated fixed defects only, nine (41%) had fixed and reversible defects and one (4.5%) had reversible defects only. Concomitant wall-thickening abnormalities occurred in 83% of segments with fixed perfusion defects, mirrored by a reduction in wall motion in 91% of segments analyzed. Right ventricular ejection fraction was correlated with age (R = 0.62; p = 0.002), and with wall-thickening abnormalities (R = 0.60; p < 0.005).
Reversible and fixed perfusion defects with concordant regional wall motion abnormalities occur in the right (systemic) ventricle 10 to 20 years after Mustard repair for transposition of the great arteries; this may be important in the pathogenesis of late right ventricular dysfunction in this group.
Transposition of the great arteries comprises 2.6% to 7.8% of all cases of congenital heart disease, and if uncorrected is commonly fatal in the first year of life (1). From the mid 1960s, operative palliation using a rerouting of systemic and pulmonary venous return (Mustard or Senning operation) was widely practiced. This procedure has a low operative mortality (2), but long-term results are compromised by progressive right ventricular enlargement and tricuspid valve regurgitation (secondary to the reliance on the right ventricle as the systemic pump) (3–5), and by atrial dysrhythmias (6). Symptomatic right ventricular dysfunction is evident in at least 10% of patients at 10 years (7).
After the atrial switch procedure, premature death has been strongly associated with the development of right ventricular impairment (8–10). While the effects of abnormal pressure loading have been studied carefully (11), little attention has been given to myocardial perfusion and the possible role of myocardial ischemia/infarction in right ventricular dysfunction (12).
We postulated that late after the Mustard/Senning operation, defects in right ventricular myocardial perfusion indicative of infarction would be demonstrable using radionuclide techniques, and that additional perfusion defects would occur during dipyridamole stress, suggesting reversible myocardial ischemia.
All surviving patients who had presented with simple d-transposition of the great arteries (d-TGA) palliated using a Mustard or Senning procedure at our center more than 10 years previously (i.e., before 1986) were eligible for the study. Patients with a ventricular septal defect or other additional cardiac defects were excluded. An irregular cardiac rhythm was also an exclusion criterion for technical reasons (gating for the nucleotide scans); patients with pacemakers were eligible if the resultant rhythm was regular. Those experiencing recent clinical deterioration (within three months) were also excluded. Our records indicated that there were 99 patients with simple d-TGA potentially suitable for the study (operation dates: 1967 to 1986), of which, 36 were under active follow-up in our center. The study protocol was discussed with 27 patients recruited consecutively at routine out-patient follow-up visits during the course of the study (9 months). Three older patients were unable to have time off from full-time employment. One was excluded with atrial fibrillation, and one was awaiting investigation of systemic venous baffle obstruction. No patient or parent refused consent. Twenty-two subjects were finally enrolled into the study.
A one-day stress/rest imaging protocol was used. Clinical evaluation and standard two-dimensional echocardiography were performed on entry to the study. Cardiac and noncardiac medications were taken as normally prescribed. After an overnight fast, each subject underwent a dipyridamole stress test with intravenous administration of (99mTc)technetium-99m sestamibi at peak stress.
Dipyridamole (0.56 mg/kg) was infused intravenously over 4 min, sestamibi injected at 6 min and aminophylline administered for the reversal of adverse effects at 8 min. The stress dose of sestamibi was 370 MBq scaled for weight. Heart rate and six-channel electrocardiogram were continuously monitored throughout the study, and a 12-lead electrocardiogram was acquired every 2 min (or more frequently if indicated). Blood pressure measurements were taken each minute.
One hour later, the stress single-photon emission computed tomography (SPECT) scan was acquired after ingestion of a glass of milk at 30 min postinjection to facilitate clearance of hepatobiliary activity. Allowing at least 4 h between scans, the second injection of sestamibi (1,110 MBq scaled for weight) was given at rest and gated SPECT imaging was performed 1 h later, again after the ingestion of milk 30 min before image acquisition.
The study protocol had the full approval of the New Children’s Hospital Ethics Committee. Written consent was obtained from the parent or subject (if over 16 years).
Demography and clinical characteristics
Table 1describes the demographic and clinical characteristics of the 22 subjects forming the study population. All but one had undergone a Mustard operation, the other a Senning, a median of 14 years previously (range 9 to 24 years). Four had had subsequent operative revision of the pulmonary venous baffle, and one had previously undergone ligation of a patent ductus arteriosus. Antiarrhythmic medication was prescribed in five subjects, two of whom were also taking an angiotensin-converting enzyme (ACE) inhibitor.
Clinical and echocardiographic assessment was performed several days before the radionuclide scan (but within one month) by an observer blinded to the results of the ongoing study. Right ventricular dysfunction-assessed echocardiographically was evaluated subjectively on a scale of 0 to 3 (0 = normal right ventricle [RV] function; 1 = mild dysfunction; 2 = moderate dysfunction; 3 = severe RV dysfunction). Time constraints resulted in no attempt being made to assess RV function quantitatively using echocardiographic techniques. Mitral and tricuspid regurgitation were assessed semiquantitatively using color-flow Doppler on a similar 0 to 3 scale (0 = no regurgitation; 1 = mild regurgitation; 2 = moderate regurgitation; 3 = severe regurgitation).
Coronary angiography and/or aortography had been performed preoperatively in most patients undergoing atrial baffle repair in our institution. Where possible, the original angiogram was examined and details of coronary artery anatomy were noted. Confirmation was sought from the original operation note where possible. Data were available in 17 of the 22 subjects (16 from the original angiogram and 1 from the operation note) and are given in Table 2.
99mTc sestamibi myocardial perfusion imaging
Single-photon emission computed tomography images were acquired using a triple-headed gamma camera (MULTISPECT 3; Siemens). High-resolution collimators were used with a noncircular orbit with two detectors undergoing a 120-degree rotation for 32 frames and 40 s per frame and a 64 × 64 matrix. Reconstruction of the stress images was done using a Butterworth filter (cut-off 0.28 to 0.40; power factor 10). The acquisition parameters were the same for the rest-gated study with the additional electrocardiogram-gated acquisition. The beat histogram was set at 20% around the peak heart rate, slightly offset to the lower side. Images were gated at eight frames per cardiac cycle using an R-wave trigger. Reconstruction of the summed ungated rest data was done using a Butterworth filter with slightly higher cut-off (0.4 to 0.5; power factor 10).
Analysis of results
Electrocardiograms obtained during the dipyridamole stress test were examined for resting and maximum heart rate, and any morphological changes occurring during or immediately after the stress test were noted.
Myocardial perfusion images were independently reviewed by three experienced nuclear medicine specialists and a consensus result obtained. The reviewers had knowledge of the underlying condition and stress test response but were blinded to the clinical status and echocardiographic data. The stress and rest perfusion studies were converted to the ICON system format for review with standard comparison views of the stress and rest images in transaxial, horizontal and vertical long-axis slices. Representative apical, mid-ventricular and transaxial slices were divided into anterior, septal, inferior and medial (free wall) segments (a total of 12 segments). Tracer uptake in each of the 12 segments was assessed subjectively as normal or reduced; reduced uptake was graded as mild (grade 1 reduction), moderate (grade 2 reduction) or severe (grade 3 reduction). Defects were confirmed on orthogonal planes (vertical and horizontal long-axis slices).
The sestamibi-gated SPECT images were read in cine mode to assess wall motion with three short-axis slices (corresponding to the same slices used for scoring the perfusion images), one mid-ventricle vertical long and one mid-ventricle horizontal long representative slice. Wall thickening and wall motion were scored subjectively as normal or reduced. In addition, for the resting gated SPECT study, comparison images of the end-diastolic and end-systolic frames were displayed in transaxial and horizontal and vertical long-axis slices. Assessment was made of wall thickening on the basis of visual assessment of brightening of the myocardial wall. Wall thickening was also graded qualitatively as normal or reduced. Right ventricular ejection fraction (RVEF) was calculated using Siemens automated gated SPECT processing software; the images were acquired during the second part of the study, representing RVEF at rest.
Descriptive data are expressed as mean ± standard deviation (range) unless otherwise stated. Correlation between variables was using standard linear regression analysis. Statistical significance was inferred at p < 0.05.
All 22 subjects enrolled in the study completed the full protocol without incident.
Dipyridamole stress test
Median resting heart rate was 70 beats per minute (range 50 to 96). An increase in heart rate was seen in all subjects during the dipyridamole infusion, with the exception of one who had a ventricular demand pacemaker. Median increase in heart rate was 28% (range 9% to 71%). PQRST morphology assessed from the standard 12-lead electrocardiogram did not show any appreciable change in any subject. Minor noncardiac effects of intravenous dipyridamole were common (as is our experience in children and adults), and were managed with intravenous aminophylline administered after the sestamibi in all but two subjects. Only one subject experienced central chest pain during the dipyridamole infusion; this settled rapidly without additional specific treatment. In no subjects were side effects sufficient to require termination of the test.
Myocardial perfusion imaging
There were 22 sets of technetium-99m sestamibi SPECT 12-segment images yielding 264 segments for analysis for both the rest scan and the stress scan. The results are summarized in Figure 1. Perfusion defects were evident in all patients in at least one segment, in either the rest or exercise images. In only one patient was the rest study normal; 55% had fixed defects only and 41% a mixture of fixed and reversible defects. Images from a representative perfusion study are reproduced in Figure 2.
Imaging at rest
At rest, perfusion defects were evident in 21 of the 22 subjects (95%) in one to six (median three) segments per subject.
Perfusion defects were observed at rest in 68 (26%) of the 264 myocardial segments analyzed. These were graded as mild (grade 1) in 47 (18%), and moderate (grade 2) in 20 (8%). There was one grade 3 defect.
Anatomically 37% of the affected segments were anterior, 18% inferior, 21% free wall and 24% septal. About a quarter of the segments were single-segment lesions, while 32% involved two adjacent segments within a single wall; nearly 40% were made up of three adjacent lesions.
Imaging during dipyridamole stress
Additional or more marked perfusion defects occurred during stress in nine subjects (45%) in one to four (median 3) segments per subject, including the one subject with a normal resting scan (three segments).
Additional or more marked perfusion defects occurred in 26 of 264 segments (10%) analyzed. Of these 26 defects, perfusion was reduced by two grades in 10 of 26 segments (38%), all of which were normal at rest, indicating a moderate reversible perfusion defect. Of the remaining 16 segments (62%), mild perfusions defects occurred in 11 segments normal at rest (reversible), and were more marked as compared with the abnormal resting scan in 5 segments (partially reversible).
Anatomically, of the 26 myocardial segments affected, 38% were anterior, 35% inferior, 3% free wall and 38% septal. Only six defects were anatomically isolated, the remainder being associated with abnormalities in one (46%) or two (34%) adjacent segments.
Prominent trabeculation was noted in the apical to mid portion of the ventricular wall in 14 subjects, seen involving the anterior wall in 13 subjects and the inferior wall in 1 subject. Small segmental perfusion defects were seen adjacent to the trabeculation in 13 of these 14 subjects. The defects were fixed in nine cases and partially reversible in four.
Right ventricular wall thickening and wall motion at rest
Two gated SPECT studies were technically inadequate and were excluded from analysis. Of the remaining 240 segments, wall thickening was abnormal in 67 segments (28%), occurring in 18 of the 20 subjects (90%).
Wall thickening abnormalities occurred in 83% (52/63) of segments with fixed perfusion defects. Of the remaining 11 fixed defects that did not demonstrate abnormal wall thickening, 5 were mild perfusion defects, single segment in size and located in the anterior wall. The six other perfusion defects (all grade 1) associated with normal wall motion occurred as three adjacent segments in the inferior wall and anterior wall, respectively, of two different patients, raising the possibility that these defects represent areas of relative attenuation rather than true perfusion defects.
Wall thickening was reduced in the absence of resting perfusion defects, in the septum of two subjects, one of whom had a reversible septal defect, and another subject showed marked global reduction in wall thickening (12 segments) in the presence of only 5 relative resting perfusion defects. No other areas of reduced wall thickening were demonstrated in sites of reversible defects which were normal at rest.
Wall motion was abnormal in 64 of 240 segments (27%), and was mirrored by reduced wall thickening in 91% of segments. In two subjects, the septum showed reduced wall motion, in the absence of resting or reversible defects. There was strong correlation between perfusion defects at rest, abnormal wall thickening and abnormal wall motion, as demonstrated in Figure 3.
Overall, fixed perfusion defects, wall motion and wall-thickening abnormalities were exactly matched in eight patients (one to six segments per patient). In a further five patients, a reduction in wall thickening was seen in the region of fixed perfusion defects in the absence of wall motion abnormality (two to three segments per patient); in two patients, the situation was reversed (three to five segments). One patient had three abnormal segments on stress perfusion imaging; wall thickening (but not wall motion) was abnormal in all three segments. In one other patient, three of four stress perfusion abnormalities were associated with both wall thickening and wall motion abnormality. With the exception noted above, no patient exhibited substantial wall motion or wall-thickening abnormalities in segments that were normally perfused during stress and at rest.
The study group was fairly typical of a post-atrial switch population with progressive deterioration of right ventricular function with age, as illustrated in Figure 4. Median RVEF was 51.5% (range 15% to 68%). The extent of wall-thickening abnormality was more strongly correlated with ejection fraction (R = 0.60; p = 0.002) than was the number of perfusion defects (R = 0.31; p = 0.16). There was a trend toward a greater number of perfusion defects in those patients with an ejection fraction of less than 40% (p = 0.12).
We have demonstrated that right ventricular perfusion defects at rest, associated with concordant wall motion abnormality, are frequent late after atrial switch repair; these defects are strongly suggestive of infarction. We have noted that significant rest defects are invariably associated with wall motion abnormality, but mild defects that are anatomically isolated may exhibit normal wall motion. Reversible defects, suggestive of stress-induced ischemia, were only occasionally associated with wall motion abnormalities at rest. However, reversible ischemia was observed in nearly half the subjects, with areas of fixed perfusion deficit, suggestive of infarction or fibrosis, associated with wall motion abnormality occurring in the vast majority. Reversibility may reflect an ongoing process and thus a window of opportunity for intervention.
Scintigraphic perfusion defects in congenital heart disease
While the value of radionuclide studies in the investigation of adults with coronary artery disease is well established, this is not the case for adults with congenital heart disease, particularly for the RV (13–15).
There is controversy over the etiology of perfusion defects detected by radionuclide techniques in children who have undergone cardiopulmonary bypass (16). Several groups have demonstrated such abnormalities in the left ventricle after the newer arterial switch operation for TGA (17,18). Weindling commented that the observed perfusion abnormalities lessened with physiologic stress in his group of 8-year-old arterial switches (19), in contrast to our own study, in which dipyridamole stress abnormalities were either constant or worsened. They concluded that the perfusion abnormalities were due to insults occurring at the time of cardiopulmonary bypass during the original operation.
A London group sought to further examine this issue in the arterial switch population by using two control groups, one having had bypass surgery without manipulation of the coronary arteries (repair of atrial septal defect, etc.), and a second group that had not had surgery at all (20). They found that perfusion abnormalities were common in both groups of operated children, but were rare in the nonoperated. All the defects noted were fixed and again tended to lessen with physiologic stress. However, no group has shown that the duration of circulatory arrest, ischemic time or bypass time are related to the incidence of such defects. Interestingly, cardiopulmonary bypass is not considered a confounding issue when interpreting myocardial perfusion scans in adults post-coronary artery surgery.
Many of the defects we have demonstrated occurred in two or more adjacent segments of the RV myocardium in contrast to the scattered pattern of defects seen in previous studies (20). Interestingly, it was the reversible defects that tended to show the most regionality (in the inferior wall and septum), suggesting limited coronary flow reserve in that area. In many cases, there were small segments of perfusion defect adjacent to areas of prominent trabeculation (associated with hypertrophy). This may possibly be due to a “steal” phenomenon and further suggest that coronary flow reserve is critically limited, particularly at high oxygen demand.
Our study differs significantly from those discussed above in the use of dipyridamole as a coronary vasodilator to assess myocardial perfusion, rather than global physiologic stress. Dipyridamole may induce two to three times greater coronary artery dilation (21) than does standard exercise, and use of dipyridamole in addition to exercise has been shown to improve diagnostic yield in adults with coronary artery disease and an inadequate exercise test (22). It is possible that dipyridamole stress is more sensitive than physiologic stress in our clinical cohort.
There is one report in the literature reporting a study similar to our own, limited to radionuclide assessment of RV perfusion at rest (23). Lubiszewska and colleagues (23) reported that perfusion defects were common (in 15 out of 26 [58%] patients) after the Mustard or Senning operation. They further concluded that older patients were more likely to have perfusion defects, although these data were confounded by older patients also having been initially operated at a later age. Lubiszewska’s patient group was much younger at the time of the study than our own (mean 10.2 years [range 5 to 20] vs. 15.5 years [range 10 to 25], respectively) with consequently shorter follow-up, yet found perfusion abnormalities that were strikingly similar. Our article extends the observations of Lubiszewska to study stress perfusion images, revealing the existence of reversible perfusion defects that might provide a focus for intervention. Furthermore, we have studied right ventricular wall motion and thickening as well as perfusion data, in the current article, to investigate the relationship between myocardial flow and function.
Since the completion of this study, our group has reported the results of a similar investigation (using treadmill exercise testing) in a small group of children and adults with congenitally corrected transposition of the great arteries (atrio-ventricular discordance and ventriculo-arterial discordance; also known as double discordance) (24). This was an unoperated group in which abnormalities of systemic right ventricular perfusion very similar to those in our study were seen, suggesting that these perfusion defects are due to right ventricular ischemia/infarction rather than to postbypass artifact. While not intended to act as a control group for the population in our current study, these data suggest that perfusion defects are not simply related to the effects of cardiopulmonary bypass.
Perfusion defects and corresponding right ventricular wall motion
To our knowledge, wall motion studies in the RV using radionuclide techniques have been reported by our group alone (24); the study design limited us to the examination of wall motion at rest. Wall motion abnormality (WMA) was demonstrated in myocardial segments associated with fixed perfusion defects, and was rare in ventricular segments with normal perfusion throughout. Wall motion abnormality in the left ventricle has been shown to correlate well with echocardiographic findings in children after the arterial switch repair (18).
Redington et al. studied RV regional wall motion angiographically in a group of patients before, and a mean of eight years after, a Mustard operation for d-TGA (25). They concluded that WMAs were common, but were often present before surgery, rarely worsening afterward. A similar conclusion regarding the progression of RV dysfunction over eight years using radionuclide ventriculography was reached by a Czech group (26). Our much longer follow-up (median 15.5 years) suggests that there is a deterioration in RV function over time. Although our population is too small to demonstrate a direct correlation of perfusion abnormalities with ejection fraction, there is good correlation of WMA with perfusion defects, suggesting an etiologic role for myocardial perfusion.
In common with other investigators, we found that profound ventricular dysfunction (and hence significant WMA) may confound interpretation of radionuclide perfusion data (27). This was evident in one subject, thought to relate to global hypoperfusion, which, given that the subject acts as his own “control” for “normal” perfusion, may result in under-reporting of hypoperfused segments.
The etiology of the observed perfusion defects is not known, but is likely to be multifactorial. The branching pattern of the major coronary vessels is often abnormal in this condition, as noted in Table 2. Furthermore, the hypertrophy of the RV when required to perform at systemic pressures for many years inevitably places a substantial demand on the coronary supply. As the coronary arterial supply of the RV is quite different from that of the left, it is possible that the right coronary arterial system may not supply adequate flow after atrial baffle repair, providing a substrate for the development of progressive ischemic damage.
The demonstration of reversible perfusion defects suggests an opportunity for treatment. Angiotensin-converting enzyme inhibitors are increasingly prescribed on a strong evidence base, in ischemic and nonischemic left ventricular dysfunction (28–30). Angiotensin-converting enzyme inhibition may represent an approach to systemic RV dysfunction, and recently reported data express cautious optimism in this regard (31), although it has been hypothesized recently that ACE inhibitors might actually be counterproductive in this situation (32).
The interpretation of right ventricular perfusion data is poorly defined in congenital heart disease. However, selective angiography, while a gold standard for anatomical assessment of the coronary circulation, does not provide functional data, while the reverse is true of echocardiography. Positron emission tomography is still undergoing full evaluation, as is magnetic resonance imaging in this circumstance (33). In contrast, sestamibi myocardial perfusion imaging allows both perfusion and function to be assessed simultaneously with minimum upset to the patient, and thus represents our best attempt yet to investigate this difficult issue.
We sought to minimize confounding issues by including only those patients with simple transposition, removing any queries relating to the interpretation of scans after ventricular septal defect repair. Additionally, assessment of wall motion in the presence of an irregular heart rhythm is difficult, and such patients were therefore excluded. Given the frequency of sinus node disease and atrial arrhythmias in this population of long-term survivors (6), only a small select group was available for study. This did, however, result in a relatively homogenous group. The coronary anatomy data were quality limited (by today’s standards) by 10- to 20-year-old technology, and related mainly to data obtained from aortography rather than from selective coronary angiography. However, other workers have considered that coronary artery anomaly is unlikely to be the cause of the demonstrated pathology (17).
Our study was too small to make any statistical comment about the influence of age on myocardial perfusion, but there was a trend toward more fixed perfusion defects in the older subjects.
Myocardial perfusion defects are common in long-term survivors of the atrial switch procedure for TGA. Our data suggest that concomitant WMA motion abnormality occurs, which is also present at rest in segments with reversible perfusion defects. It is probable that these findings represent areas of infarction and ischemia within the RV, and may contribute to late RV dysfunction in these patients. Further prospective studies might assess whether interventions designed to reduce ischemia may slow the progression of right ventricular impairment in this group of young adult survivors of congenital heart surgery.
- angiotension converting enzyme
- d-transposition of the great arteries
- right ventricle
- right ventricular ejection fraction
- single-photon emission computed tomography
- wall motion abnormality
- Received July 22, 1999.
- Revision received October 5, 1999.
- Accepted January 12, 2000.
- American College of Cardiology
- Liebman J.,
- Cullum L.,
- Belloc N.B.
- Kirklin J.,
- Blackstone E.,
- Tchervenkov C.,
- Castaneda A.
- ↵Gellatt M, Hamilton R, BW M, et al. Arrhythmia and mortality after the Mustard procedure: a 30 year single centre experience. J Am Coll Cardiol 1997;29:194–201.
- Martin R.,
- Shakeel Q.,
- Ettedgui J.,
- et al.
- Gewillig M.,
- Cullen S.,
- Mertens B.,
- Lesaffre E.,
- Deanfield J.
- Redington A.,
- Rigby M.,
- Shinebourne E.,
- Oldershaw P.
- Flynn B.,
- Wernovsky G.,
- Summerville D.,
- Castaneda A.,
- Treves S.
- Bajawa N.,
- Bianco J.,
- Stone C.
- Imbriaco M.,
- Cuocolo A.,
- Pace L.,
- et al.
- Vogel M.,
- Smallhorn J.,
- Gilday D.,
- et al.
- Weindling S.,
- Wernovsky G.,
- Coaln S.,
- et al.
- Hayes A.,
- Baker E.,
- Kakadeker A.,
- et al.
- Candell-Riera J.,
- Santana-Boado C.,
- Castell-Conesa J.,
- et al.
- Lubiszewska B.,
- Gosiewska-Marcinkowska E.,
- Teresinska A.,
- Rozanski J.,
- Rydlewska-Sadowska W.,
- Ruzyllo W.
- Hornung T.,
- Bernard E.,
- Jaeggi E.,
- Howman-Giles R.,
- Celermajer D.,
- Hawker R.
- Redington A.,
- Rigby M.,
- Oldershaw P.,
- Gibson D.,
- Shinebourne E.
- Reich O.,
- Voriskova M.,
- Ruth C.,
- et al.
- Chua T.,
- Kiat H.,
- Germano G.,
- et al.
- ISIS 4 Pilot Study Group
- ↵Lubiszewska B, Gosiewska E, Teresinska A, Ruzyllo W. Myocardial perfusion of the systemic right ventricle in patients after the atrial switch for complete transposition (TGA): long-term follow-up. The Second World Congress of Pediatric Cardiology and Cardiac Surgery, Honolulu, HI; 1997:146 (abstr).
- Deanfield J.E.