OUP user menu

Decline in ventricular function and clinical condition after mustard repair for transposition of the great arteries (a prospective study of 22–29 years)

J.W Roos-Hesselink, F.J Meijboom, S.E.C Spitaels, R van Domburg, E.H.M van Rijen, E.M.W.J Utens, J McGhie, E Bos, A.J.J.C Bogers, M.L Simoons
DOI: http://dx.doi.org/10.1016/j.ehj.2004.03.009 1264-1270 First published online: 2 July 2004


Background Great concern exists about the ability of the anatomic right ventricle to sustain the systemic circulation in patients with transposition of the great arteries who have undergone a Mustard procedure. A prospective study was made to examine long-term survival, clinical outcome, and right ventricular function 25 years after surgery.

Methods Ninety-one consecutive patients underwent the Mustard procedure between 1973 and 1980. After 14 years and again after 25 years (range 22–29 years), patients were studied with ECG, echocardiography, exercise testing, and Holter monitoring.

Results The cumulative survival and event-free survival were 77% and 36%, respectively, after 25 years. Reoperation was necessary in 46%. No major loss of sinus rhythm was found.

While all patients had good right ventricular function 14 years after repair, 61% of patients showed moderate-to-severe dysfunction after 25 years, when studied by echocardiography. Furthermore, the QRS complex widened and exercise capacity decreased.

Conclusion The anatomic right ventricle appears to be unable to sustain the systemic circulation at long-term follow-up and the clinical condition of patients late after Mustard repair is declining. We can expect more deaths or need for heart transplantation in the next decade.

  • Transposition of great arteries
  • Congenital heart defects
  • Survival
  • Follow-up studies
  • Echocardiography


The atrial switch procedure dramatically improved the survival of children born with complete transposition of the great arteries. Major concerns for long-term outcome are: the fate of the right ventricle supporting systemic circulation, arrythmias, and, accordingly, the long-term survival of these patients. 1–5 Systemic right ventricular dysfunction and loss of sinus rhythm have been reported at long-term follow-up, but it is unclear whether these problems were found in a selection of “worse cases” or whether systemic right ventricular failure and loss of sinus rhythm are inevitable in all patients. 3,5–8 We made a systematic follow-up study of a single-centre cohort of 91 consecutive patients who had undergone the Mustard procedure in the 1970s. This cohort was investigated in 1990. The outcome then showed good right ventricular function in most patients, with the majority of patients in sinus rhythm (69%). In 2001, 11 years after the first study and 22–29 years after the Mustard operation, this same cohort was studied again to assess survival, clinical course, arrhythmias, and systemic ventricular function and the findings were compared with the data from the study of 1990.



Ninety-one consecutive patients who underwent Mustard repair for transposition of the great arteries at the Thoraxcenter between 1973 and 1980, who were younger than 15 years at the time of surgery, formed the study group of this report. Before the Mustard operation, improved atrial mixing was achieved in 15 of the 91 patients (16%) by Blalock–Hanlon atrioseptostomy, and in 65 of the 91 (71%) by Rashkind balloon atrioseptostomy. Operations were performed with mild hypothermia and complete cardiopulmonary bypass in 18 patients (all between 1973 and 1976), and with deep hypothermia and circulatory arrest in 64 patients. In nine patients the cooling technique was not described in detail. Cold cardioplegia was used in 38 patients (42%) after 1976. In 36 patients, additional surgery for ventricular septal defect (VSD) closure and/or pulmonary valve stenosis (PS) was performed. These 36 patients are henceforth referred to as “complex Mustard”.


Survival status was established in all 91 patients. In order to assess cardiac function, a first systematic follow-up study was made in 1990. At that time, 18 patients had died and five had moved abroad. Fifty-eight of the remaining 68 patients (79% of the survivors, 85% of those eligible for follow-up) agreed to participate in the first follow-up study. These 58 patients were invited for a second follow-up in 2001. Between 1990 and 2001, two more patients died, one patient underwent cardiac transplantation, and one patient refused to participate, so 54 participated in the second follow-up study in 2001. Patients who did not participate in this study were contacted by telephone. The evaluations in 1990 and 2001 consisted of a detailed history, physical examination, standard electrocardiogram (ECG), 24-h ambulatory ECG, exercise test, and echocardiogram. The institutional Medical Ethics Review Board approved the study. All patients gave their consent.


Standard 12-lead surface electrocardiograms were analysed for rhythm, P-wave axis, P-wave height and duration (measured in lead II), PR interval, and QRS duration. A QRS duration of more than 120 ms was defined as complete bundle-branch block. A single observer made all ECG measurements.

Twenty-four-hour ambulatory ECG

A three-channel recorder was used. Sinus-node dysfunction was assessed using the modified Kugler criteria and was judged to be present in the case of nodal escape rhythm, sinus arrest Math3 s, or severe sinus bradycardia (Math30 beats/min at night or Math40 beats/min during the day).9


Transthoracic two-dimensional echocardiography and echo-Doppler studies were made using a Hewlett-Packard Sonos 5500 echocardiograph. The same echocardiographer performed a systematic echo assessment in 1990 and 2001. Baffle obstruction was diagnosed if there was evidence of obstruction in both the two-dimensional image and Doppler flow pattern: low-flow velocity (Math0.5 m/s) in the superior or inferior caval vein and acceleration to Math1.0 m/s in the baffle with the typical, flat, non-pulsating curve profile. Obstruction to pulmonary venous drainage was diagnosed if the peak velocity at the junction of the pulmonary venous atrium and right atrium exceeded 1.5 m/s. A systematic approach to evaluating right ventricular dimensions and function was applied separately by two experienced cardiologists blinded for the other data. Interobserver variability was analysed, and a kappa value of Math0.7 was considered excellent. Dimensions were judged as either normal or mildly, moderately, or severely dilated. Right ventricular systolic function was graded as either normal or mildly, moderately, or severely impaired. Acoustic quantification is an automatic border detection technique that has been used for the quantitative assessment of ventricular volume and performance and was used to assess the right ventricular ejection fraction in patients where it was technically feasible. An ejection fraction of Math50% was defined as normal. All echocardiographic studies of 1990 were re-evaluated. The degree of tricuspid regurgitation (minimal, moderate, or severe) was estimated with colour-Doppler by the width and length of the regurgitant jet and the Doppler flow pattern in the pulmonary veins. Pulmonary hypertension was defined by an early diastolic pulmonary regurgitation flow velocity of Math2.5 m/s or, in the absence of left ventricular outflow tract obstruction, a mitral regurgitation flow velocity of Math3.0 m/s.

Maximal exercise capacity

Maximal exercise capacity was assessed by bicycle ergometry with stepwise increments of 20 W of workload per minute. Exercise capacity was compared to that of normal individuals, corrected for age, sex, and body height. Exercise capacity Math85% of the predicted normal value was considered abnormal.

Major events were defined as death, cardiac transplantation, reoperation, pacemaker implantation, or hospital admission for arrhythmia, endocarditis, or heart failure.

Data analysis

Data are presented as median and range, unless indicated otherwise. The Math test and Fisher's exact test were used to compare discrete variables. The Student's t test was used to compare continuous variables. All tests were two-tailed. The McNemar test of symmetry was used to compare the 14-year and 25-year outcomes. Cumulative survival curves were constructed using the Kaplan–Meier method. Among patient subgroups, the log rank test was used to compare survival curves. The level of significance was Math. Multivariable Cox-regression analysis was performed for survival, event-free survival, and right ventricular dysfunction. The variables tested were birth weight, type of palliation, “complex” Mustard procedure, age at operation, year of operation, use of deep hypothermia, use of cold cardioplegia, aortic cross-clamp time, right ventricular incision, perioperative complications, residual lesions, and rhythm at first outpatient visit after operation. Continuous variables were not categorised in the model. The proportional hazards assumptions were tested by constructing interaction terms between the variables and time to each endpoint. Cox regression analyses showed no statistically significant interactions with time (each Math). The model selection is based on the stepwise principle, where the limits for entering and removing variables were both 0.05.


Median age when the Mustard operation was performed was 2 years (range 0–11 years). Further baseline characteristics and surgical details have been reported previously.6 Information on survival is complete for the total cohort of 91 patients.


The cumulative survival after Mustard repair was 80% after 14 years and 77% after 25 years (Fig. 1). Twenty patients died, five of them in the perioperative period. Of the 15 patients with late deaths (Math30 days after surgery), three had right ventricular failure and pulmonary oedema diagnosed directly after the Mustard procedure, 3 others had atrial flutter/fibrillation, seven died suddenly 1–15 years after surgery, and two patients died 23 and 26 years after surgery due to progressive right ventricular pump failure. One patient underwent successful heart transplantation 26 years after her Mustard repair because of failure of the systemic right ventricle. We found no differences in survival between patients with simple versus complex Mustard operation (Fig. 2).

Fig. 1

Survival (event-free) after Mustard operation. Crossbars mark the 95% confidence intervals.

Fig. 2

Survival after the Mustard operation for transposition of the great arteries.

Major events

Cumulative event-free survival after 29 years was 36% (Fig. 1). Since the initial Mustard procedure, 25 patients (46%) have been reoperated (Table 1). One more patient had successful balloon angioplasty of a baffle obstruction. The median interval between the Mustard repair and reoperation was 11 years (range 2–29 years). No patients died during reoperation.

View this table:
Table 1


NumberTime since initial repair (+range) in years
Baffle stenosis1713 (2–23)
Arterial switch29 (8–9)
Residual VSD29 (2–16)
Aortic coarctation18
Patent arterial duct12
Subpulmonary stenosis16
Tricuspid and mitral valve111

Of the patients seen at the last follow-up, 5 (9%) were admitted to the hospital for heart failure. Pacemaker implantation was performed in 16 patients (28%), four of them in the period between 14 and 25 years of follow-up, all four for sinus node disease. Between 14 and 25 years of follow-up, three patients (5%) were admitted for electric cardioversion of supraventricular tachycardia and two for endocarditis. A total of 42 patients (61%) had at least one hospital admission (for heart failure, endocarditis, arrhythmia, or reoperation) in the total follow-up period.

Clinical evaluation

The median age at evaluation was 16 years in 1990 and 27 years (range 22–38 years) in 2001, with a median follow-up of 14 years and 25 years (22–29 years), respectively, after surgery. The clinical condition worsened between 1990 and 2001. After 25 years of follow-up, 24% of the patients were in New York Heart Association (NYHA) class I (Fig. 3). Fifteen patients (29%) were taking one or more drugs: oral anticoagulants in two, β-blockers in two, digoxin in three, antiarrhythmics in four, and angiotensin-converting enzyme (ACE) inhibitors in 10 patients. ACE inhibitors were given for evident heart failure in two patients and for right ventricular dysfunction in eight patients. Of the 15 female patients, four had successful pregnancies.

Fig. 3

NYHA classification in 1990 and 2001.

Mean oxygen saturation was 97% (range 80–100%). Signs of heart failure (oedema, liver enlargement, and elevated central venous pressure) were found at last follow-up in one patient.


The ECG findings are presented in Table 2. The percentage of patients with sinus rhythm on the 12-lead electrocardiogram diminished from 69% in 1990 to 63% in 2001. The PR interval and duration of the P wave increased, but these changes were not statistically significant. The increase in QRS duration from 94 to 110 ms was highly significant (Math) (Fig. 4), as was the change in the QRS-axis. QTc duration did not change.

View this table:
Table 2

Standard 12-lead ECG and 24-h Holter electrocardiogram

14 years25 yearsFormula value
12-lead ECG
Sinus40 (69%)34 (63%)
Nodal7 (12%)7 (13%)
Atrial6 (10%)5 (9%)
Pacemaker4 (7%)8 (15%)
Aflutter1 (2%)
PR interval (ms)1571650.09
QRS duration±SD (ms)95±11110±17Formula0.0001
QTc segment±SD (ms)395±32396±400.7
QRS axis±SD142±56155±600.008
P-wave duration±SD86±1391±150.06
P-wave height±SD0.25±0.140.25±0.140.9
P-wave axis±SD58±2359±260.8
24-h Holter
Sinus node disease37%43%Formula0.001
Paroxysmal A fib/flutter4%00.2
VT Formula10 beats00
VT 3-10 beats7%8%0.7
Fig. 4

QRS-prolongation between 1990 and 2001.

At last follow-up, one patient had first-degree atrioventricular block and 9 patients had bundle-branch block.

Twenty-four-hour ambulatory monitoring

At last follow-up, none of the patients had atrial flutter or fibrillation (see Table 2). Junctional escape rhythm was seen in 23%. Signs of sinus node disease increased significantly between 14 and 25 years of follow-up, but no ventricular pauses longer than 3 s occurred. Ventricular tachycardia of more than 10 complexes was not observed; 8% showed ventricular tachycardia of 3–10 complexes.


The right (systemic) ventricle was dilated in 72% of the patients, compared to 58% ten years earlier. The function of this systemic ventricle changed markedly (Fig. 5). While right (systemic) ventricular function was normal in 69% after 14 years of follow-up, ten years later only 6% of the patients had normal ventricular function (Math). Ventricular function deteriorated from good to mildly (33%), moderately (37%), and severely dysfunctional (24%). The interobserver variability was excellent, with 90% agreement and a kappa value of 0.86. In 10 patients it was possible to measure the ejection fraction by acoustic quantification, which ranged from 19% to 53% with a mean of 38%. Tricuspid regurgitation was severe in only 2% of patients in 1990, compared to 20% in 2001. Tricuspid regurgitation was not more frequent in patients who had undergone a “complex” Mustard procedure. Pulmonary hypertension was found in 3% in 1990 and in 18% in 2001. Twenty-three patients had 30 residual lesions: baffle obstruction in 17, baffle leakage in two, residual VSD in three, and pulmonary valve stenosis in eight patients.

Fig. 5

Decline in systemic ventricular function between 1990 and 2001 measured with echocardiography.

Exercise tolerance

Fifty patients exercised to maximal effort (1 patient was excluded because of spastic hemiplegia, one had psychomotor retardation, and two patients refused). The maximal exercise capacity (Fig. 6) in these patients was 72% (range 38–105) of the predicted normal values in 2001, compared to 84% in 1990 (Math). Thirty-eight patients (76%) had an exercise capacity below 85% of predicted value. The maximal heart rate was 170 beats/min (82% of expected) in 1990 and 165 beats/min (84% of expected) in 2001. During the test, five patients (10%) developed arrhythmias: three patients had an increase in ventricular extrasystoles and two patients developed paroxysmal supraventricular tachycardia. No ventricular tachycardia occurred.

Fig. 6

Decrease in exercise capacity between 1990 and 2001.

Risk factors for late-death major events and systemic ventricular dysfunction

Multivariable Cox-regression revealed one independent predictor for death, which was the period when surgery was performed: patients operated between 1973 and 1976 showed a worse survival than those operated between 1977 and 1980 (Table 3). For the endpoint “moderate-to-severe dysfunction of the right ventricle at last follow-up”, two independent predictors were found: atrial flutter at the first outpatient visit after Mustard repair and additional repair of VSD and/or PS (complex Mustard procedure).

View this table:
Table 3

Multivariable Cox regression

Independent predictorsHR95% CI
Endpoint: death
Surgery performed before 19773.11.1–9.5
Endpoint: moderate to severe right ventricular dysfunction
Atrial flutter at first follow-up after Mustard repair3.01.1–8.9
Complex Mustard procedure (VSD and/or PS)1.91.1–4.0


This series is unique in that it comprises a cohort of consecutive patients who underwent surgery at a single institution and had a longitudinal follow-up of 22–29 years or until death. Our study shows some late attrition and substantial long-term morbidity after the Mustard operation. Most importantly, a clear deterioration in systemic right ventricular function and clinical condition was seen in virtually all patients from 14 to 2522–29 years after Mustard repair for transposition of the great arteries.


The cumulative survival was 80% after 14 years and 77% after 25 years. We found a small, continuing attrition rate as the population aged. In contrast with other authors, we did not find a difference in survival rate between patients with simple or complex Mustard procedure.10,11 In the literature, the most frequent cause of late mortality after the Mustard procedure is sudden death, presumably arrhythmic.2,10–12 In the present study, the major cause of death changed from sudden, or arrhythmic, in the first 14 years after surgery, to progressive heart failure in the period from 14 to 25 years after surgery.

Morbidity was substantial in our Mustard population, with an event-free survival 25 years after surgery of only 36%. Reinterventions were required in nearly half of the patients. Of the patients seen at last follow-up, 28% needed a permanent pacemaker and 29% were using cardiac medication.

Assessment of right ventricular function

Although echocardiographic analysis of systemic right ventricular function has its limitations, we chose to use this technique again in 2001 to enable comparison with the 1990 study. With magnetic resonance imaging (MRI), 28% of the patients would have been excluded because of a permanent pacemaker and we would not have been able to compare the data with the study of 1990.

We found an excellent interobserver variability on visual estimation of right ventricular function.

Systemic right ventricular function

We observed a striking deterioration of systemic right ventricular function in the decade from 14 to 25 years of follow-up. While all patients had good function or only mild dysfunction 14 years after repair, after a median follow-up of 25 years, moderate-to-severe dysfunction developed in 61% of patients, mild dysfunction in 33%, and only 6% conserved good function. This time course suggests that although the right ventricle can sustain the systemic circulation for about two decades, development of right ventricular failure is inevitable in the long run. Another expression of systemic ventricular failure is the increase in the incidence of pulmonary hypertension from 3% to 18% and the increase of severe tricuspid regurgitation from 2% to 20%.

We found a rightward shift of the QRS axis and marked widening of the QRS complex. The increase in QRS duration may be another indicator of deterioration of right ventricular function in these patients.13 Patients' exercise capacity decreased from 84% of the predicted value in 1990 to 72% in 2001 (Math). No significant change in maximal heart rate with exercise was found. The cause of the reduction in exercise capacity is still a matter of debate and can be due to impaired systemic ventricular dysfunction, chronotropic incompetence, or the fixed atrioventricular filling rate caused by the baffle situation. Chronotropic incompetence was not found in our study. Although 77% of the patients still are in fairly good clinical condition (NYHA I or II) after 22–29 years, we saw a deterioration in functional class in 10 years. Thus, the physical condition of the post-Mustard patients is declining. Echocardiographic data, electrocardiographic measurements, exercise capacity, and functional class all show signs of deterioration between 14 years and 25 years after surgery. These consistent findings were observed in nearly all patients. Ventricular dysfunction is recognised as a risk factor for heart failure and death in patients with a systemic right ventricle.3 If this deterioration progresses at a similar pace, it is likely that many of these patients, now aged 20–38 years, will develop heart failure, die, or need heart transplantation in the next decade. In patients with congenitally corrected transposition, with the right ventricle supporting the systemic circulation, a survival rate of only 20% at age 60 is found and we may find this same dreadful figure for the Mustard patients as well.14 Until now, the expectation was that the Mustard patients would have a better life expectancy, but the development of systemic ventricular dysfunction shows otherwise. The median follow-up in our study (25 years) is the longest reported in the literature and we are now beginning to see that the outcome of Mustard patients is, in fact, comparable to the outcome of patients with congenitally corrected transposition. 2,5,8,15–18 Whether or not medication will improve this outcome needs to be investigated. Because abnormal ventricular filling may be due, in part, to flow limitations originated by the atrial baffle, the use of ACE inhibitors may be counterproductive; no positive results have been found until now.19 The late arterial switch operation has not shown good results in adults so far and therefore is not an attractive option either.20,21 Whether resynchronisation therapy with pacing modalities will be of value in these patients, as it seems to be in patients with systemic left ventricular dysfunction and wide QRS complexes, needs to be investigated.22

Cause of systemic right ventricular dysfunction. The cause of right (systemic) ventricular dysfunction remains uncertain. Impaired atrioventricular transport, secondary to the rigid baffle construction, may be a contributing factor.23–25 In our study, however, baffle obstruction occurring in 1990 was not found to be a predictor of right ventricular dysfunction in 2001. We could identify some other risk factors for ventricular dysfunction, such as atrial flutter at first follow-up after operation and complex versus simple Mustard procedures. The atrial arrhythmias that appear shortly after surgery are probably an expression of the impaired haemodynamic situation, as described by others.26 Millane found myocardial perfusion defects, and Singh found impaired myocardial flow reserve in the systemic right ventricle in survivors of the Mustard operation, suggesting that coronary blood supply is the main problem.27,28 Lubiszewska29 confirmed these findings and found greater perfusion defects more frequently in older patients with longer follow-up, which is in line with our study of deterioration of ventricular function over time. The fact that almost all patients develop right systemic ventricular dysfunction supports the theory that right ventricular dysfunction is inherent to the Mustard situation itself, with the anatomic right ventricle functioning as the systemic ventricle.

Atrial arrhythmias

Although early atrial arrhythmias were a predictor for late ventricular dysfunction in our study, arrhythmias occurring late after surgery could not be identified as a risk factor for right ventricular dysfunction or death, as has been reported elsewhere.2,18 We found no significant ongoing loss of sinus rhythm and the incidence of atrial fibrillation and flutter was low in the last 10 years, with only three patients needing electric cardioversion. Furthermore, no case of atrial flutter or fibrillation was found on electrocardiograms, 24-h Holter monitoring, or exercise testing. Sinus node disease was the most common arrhythmia but most patients received their pacemaker in the first years after surgery and only four patients have required pacemaker implantation in the last 10 years.

Study limitations

The number of patients in this study is limited, as in other studies of adult congenital heart disease. However, we report the follow-up of a consecutive cohort of operated patients, with no selection bias related to disease severity.

It should also be appreciated that the assessment of right ventricular function by echocardiography has its limitations. MRI may serve as an additional method for studying right ventricular function. However, in patients who have a pacemaker, magnetic resonance is not a suitable technique and MRI was not available in 1990. In the present study, serial echocardiography was chosen to compare data over time, and each patient was used as a personal control.


We thank Mrs. V.E. Kleyburg-Linkers (echocardiography laboratory), Mrs. E.M. Peterse-Dekkers (Holter laboratory) for her excellent technical support, and Mrs. W. van der Bent for her help in preparing the manuscript.


  • The report was written as part of a project funded by the Dutch Heart Foundation (No. 99.033).


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
View Abstract