Download PDF

Between 1988 and 1994, the mortality rate for the arterial switch operation (ASO) at Bristol Royal Infirmary (BRI) was about double that in England as a whole. 1,2 The program was therefore suspended in 1995 and a public inquiry, under the chairmanship of Professor Ian Kennedy, concluded that “the Bristol phenomenon reflected a malaise within the National Health Service itself, with individuals who failed to work together effectively for the interests of patients, and units lacking in leadership and teamwork”. 1,2 He called for a more open and accountable health service in which patients were seen as partners in decision-making, standards for clinical and hospital practice were set and monitored, and local performances were measured and published.

At the same time a major reorganization of the paediatric cardiac service in Bristol took place. All open and closed paediatric cardiac surgical procedures were moved to the children’s hospital rather than a split site ar- rangement with open heart surgery in children being delivered within an adult cardiac surgical service at the BRI. A new surgeon and paediatric cardiac anaesthetic team were appointed and the paediatric intensive care unit developed to accommodate the needs of infants and children during their post-operative recovery. At the same time, a direct link with the Bristol Heart Institute was established to start a research program aiming at monitoring and improving clinical outcomes.35

Today the Bristol Royal Hospital for Children (BRHC) provides a comprehensive congenital cardiac surgical service to the South West of England and South Wales covering a population of approximately 4.5 million. Since the “Bristol scandal” and the changes implemented in 1995 more than 2800 congenital heart operations have been performed of which around 50% were on patients below 1 year of age.6

In this report we sought to analyze our 10 year experience with the ASO in the current era, with an emphasis on early and long term surgical results. Furthermore we also present the experience of mentoring a newly ap- pointed paediatric surgeon to perform the ASO independently.


This study analyses data for patients who underwent ASO between November 1995 and September 2007. Data came from three sources: (a) an institutional database of preoperative characteristics, surgical information, in-hospital outcomes and follow up collected prospectively on all cardiac surgery patients treated in our unit; (b) death registration from the UK National Strategic Tracing Service (NSTS); (c) review of patient hospital notes.

During the study period, 138 ASOs were performed by three surgeons (AP since 1996, AJP since 2000 and MC since 2005) at the BRHC. Of these, 41 patients (30%) were classified as complex TGA because of the presence of concomitant ventricular septal defetct (VSD), aortic coarctation or hypoplastic aortic arch and Taussig-Bing form of double outlet right ventricle (DORV) with sub-pulmonary VSD.Table 1 shows the preop-erative patient characteristics. Previous operations prior to ASO (pulmonary artery banding with aortopulmonary shunt) were performed in 2 patients.

Table. 01

Baseline characteristics.

Number of ASO 138
Age (days) 11 (8-15)
Weight (kgs) 3.4 (3.1-3.7)
Complex TGA 41 (30%)
VSDs 30 (21.7%)
Aortic coarcation or hypoplasticarch 9 (6.5 %)
TGA+DORV 6 (4.3%)
Unusual coronary anatomy 34 (25%)
Type D 19 (14.1%)
Type E 7 (5.2%)
Intramural 4 (3.0%)
Single coronary artery 5 (3.7%)
Unmatched echocardiographic diagnosis for unusual coronary anatomy 6 (17%)
Pre operative septostomy 100 (74%)

Note. ASO=arterial switch operation; VSD=ventricular septal defect; TGA=transposition of the great arteries; DORV=double outlet right ventricle.

Anaesthetic and Surgical Management:

Since the changes had been put in place at the BRHC, new facilities and protocols for children born with trans- position of the great arteries (TGA) have been instituted. New facilities included a dedicated paediatric cardiac theatre and paediatric intensive care unit (PICU) located at the new Bristol Royal Hospital for Children. A new surgical and anaesthetic team lead by AP and AW restarted the switch program at the beginning of 1995 initially with the cooperation of staff from the Birmingham Children’s Hospital.

The anaesthetic team consisted of four dedicated paediatric anaesthetists solely committed to paediatric cardiacanaesthesia and intensive care. Formal check lists were instituted with computer generated individualised drug charts, based on weight and surface area, to ensure that drugs could be prepared rapidly even by junior mem-bers of the anaesthetic and intensive care team.

Surgery was performed using systemic hypothermia (18-24 C) and low flow. Complete circulatory arrest was used selectively. Cold crystalloid cardioplegia which was used initially, was replaced by cold blood cardiople-gia (4:1 dilution blood/St Thomas’ I crystalloid cardioplegia) after several reports from our Unit demonstrating improved myocardial metabolic preservation and protection with this technique.7,8 In all but 2 patients with side-by-side great vessels, the Lecompte maneuver was used to bring the pulmonary artery anterior to the aorta. Coronary transfer was undertaken using the trap door technique in all patients.9 For pulmonary artery reconstruction the “pantaloon patch” technique, introduced by Paillole et al.,10 was used. Modified ultrafiltration at the end of cardiopulmonary bypass was used in the majority of patients (96%).

Postoperative Management and Assessment of Clinical Outcome:

All patients were admitted to PICU after surgery and managed by paediatric intensivists and cardiologists. Decisions regarding inotropic support and ventilation were based on unit protocols, haemodynamic status (e.g.mixed venous saturation, lactic acidosis, post-operative echocardiographic data) and clinical judgment.

Intraoperative and postoperative clinical parameters were measured. These included the duration of CPB, aortic cross-clamp, difficulties in coming off CPB for haemodynamic instability, inotropic support, duration of ventilation, ICU and hospital stay. In-hospital death was defined as any death occurring during the same hospital admission as surgery. Post-operative complications were defined as minor (wound and chest infections, transient renal or respiratory insufficiency, pleural effusion, transient arrhythmias or pace maker requirements) or major (any cardiac conditions requiring re-intervention, myocardial infarction, permanent stroke, diaphragmatic paralysis requiring surgical plication, septicaemia, permanent heart block).

Postoperative follow-up data, from 136 patients (100% of the survivors), were collected from the outpatient records. Besides clinical follow-up, electrocardiogram (ECG), x-ray and two-dimensional echocardiography with Doppler flow studies were performed at 6-month intervals during the first postoperative year and annually thereafter. Major follow up events were defined as death, myocardial infarction, deterioration of ventricular function assessed by echocardiography, and any surgical or catheter-based procedure.

Statistical Analysis:

Data are reported as number and percentage (categorical data) or median and interquartile range (IR) for continuous data. Baseline characteristics and outcomes were compared using the chi-squared or Fisher’s exact test (categorical variables) or the Wilcoxon rank sum test (continuous variables). Survival curves were constructed using the Kaplan-Meier method. A p- value less than 0.05 was considered to be statistically significant.


Early Mortality and Morbidity:

Early mortality occurred in 2 patients (1.5%) with complex TGA. One patient with diagnosis of TGA, VSD and aortic coarctation died of progressive myocardial failure and sepsis 24 days post-operatively. The other patient with TGA and VSD developed fatal myocardial infarction 1 day after switch operation related to an intramural course of the coronary artery. As expected, simple TGA was associated with better early outcomes compared with complex TGA (Table 2). Simple TGA required shorter cross clamp and CPB time, less problems in coming off CPB, less major postoperative complications and consequently shorter ICU and hospital stay.

Table. 02

Operative characteristics and clinical outcomes.

Outcomes All TGA n=138 Simple TGA n= 97 (70%) Complex TGA n=41 (30%) P value
In-hospital mortality 2 (1.4%) 0 2 (4.8%) 0.02
Cross-clamp time(min) 65 (51-62) 60 (51-83) 74 (62-86) 0.001
CPB time (min) 110 (87-138) 104 (86-136) 120 (100-162) 0.004
Lowest body temperature (°C) 22 (18-22) 22 (20-22) 22 (18-22) 0.1
Modified Ultrafiltration 125 (92%) 87 (91%) 39 (95%) 0.3
Difficult weaning from CPB 7 (5%) 2 (2.0%) 5 (12.1%) 0.01
Ventilation time (hours) 48 (36-72) 48 (24-72) 49 (48-72) 0.05
ICU stay (days) 4 (3-7) 4 (3-7) 4 (3-9) 0.1
Post-operative complications 0.01
Major 10 (7.3%) 4 (4.1%) 6 (14.6%)
Minor 13 (9.5%) 9 (9.2%) 4 (9.7%)
Hospital stay (days) 13.5 (10-20) 13 (10-19) 15.5 (12-21) 0.02

Note. CPB=cardiopulmonary bypass; TGA=transposition of the great arteries; ICU=intensive care unit.

Follow up:

Late death was observed in 2 patients within 6 months after the ASO. One patient with an intramural left coronary artery underwent the insertion of a 5 mm Goretex tube to reconnect the coronary ostium to the ascending aorta because of surgical damage to the coronary artery. He was discharged home in stable haemodynamic condition on full anti platelet therapy, but was readmitted 4 months later with ischaemic ECG changes due to sudden occlusion of the Goretex shunt and died of myocardial infarction. The second patient with a diagnosis of TGA and multiple VSD with Taussig-Bing complex underwent ASO and VSD closure followed by pulmonary artery banding for a residual VSD. He was readmitted 3 months later with heart failure and ventricular dysfunction and an angiogram showed coronary malperfusion. The patient underwent emergency reoperation with refashioning of the aortic root and VSD closure but died subsequently of myocardial dysfunction. Figure 1 shows the survival and the cardiac event free survival for the simple and complex TGA. At a follow up of 5 years simple TGA (n=97) undergoing ASO showed a significantly better survival and cardiac event free survival compared to complex TGA (n=41) [98.9% (CI 1.0 to 1.9) and 92.7% (CI 4.1 to 8.0), p=0.04] and [92.8% (CI 2.9 to 5.7) and 79.9% (CI 6.4 to 12.5), p=0.03] respectively. The overall survival and cardiac event free survival for the entire group (n=138) at a mean follow up of 5.8±3.2 years was 97.1% (CI 0.9 to 6.8) and 89% (CI 6.4 to 16.9) respectively (Figure 2).

Kaplan Meier curves, showing the observed cumulative percentage of patients with simple (0) or complex (1) TGA who survived with increasing duration from the time of the ASO. Kaplan Meier curves, showing the observed cumulative percentage of patients with simple (0) or complex (1) TGA who survived or did not experience one or more cardiac-related event with increasing duration from the time of the ASO.

Figure. 01A Kaplan Meier curves, showing the observed cumulative percentage of patients with simple (0) or complex (1) TGA who survived with increasing duration from the time of the ASO. Figure. 01B Kaplan Meier curves, showing the observed cumulative percentage of patients with simple (0) or complex (1) TGA who survived or did not experience one or more cardiac-related event with increasing duration from the time of the ASO.

Figure. 02A Kaplan Meier curves, showing the observed cumulative percentage of patients with TGA who survived with increasing duration from the time of the ASO. Figure. 02B Kaplan Meier curves, showing the observed cumulative percentage of patients with TGA who survived or did not experience one or more cardiac-related event with increasing duration from the time of the ASO.

Figure. 02A Kaplan Meier curves, showing the observed cumulative percentage of patients with TGA who survived with increasing duration from the time of the ASO. Figure. 02B Kaplan Meier curves, showing the observed cumulative percentage of patients with TGA who survived or did not experience one or more cardiac-related event with increasing duration from the time of the ASO.

Cardiac events at follow up included balloon dilatation or stenting of stenosed pulmonary arteries (n=7; 5.2%), reparation for pulmonary artery reconstruction (n=1; 0.7%), heart transplantation for progressive dilated cardiomyopathy but no angiographic evidences of coronary insufficiency (n=1; 0.7%), residual ASD closure with device (n=1; 0.7%) and ballooning of a recurrent narrowing at the site of the aortic arch repair (n=1; 0.7%).

Training a Paediatric Cardiac Surgeon:

In 2005 a new Paediatric Consultant Cardiac surgeon (MC) joined the team of 2 experienced surgeons (AP and AJP). He had previously been trained as a registrar in paediatric cardiac surgery and was already an independent adult cardiac surgeon. Over the previous 2 years he had assisted in 33 ASO and had became an independent paediatric cardiac operator for less complex congenital heart procedures in infants and children. The training process for ASO, following observation and assisting, consisted of the performance of part of the switch operation (such as the reconstruction of the right ventricular outflow tract) progressing to the harvesting of the coronary sinuses, followed by performing the whole ASO procedure. The mentoring process continued with 5 ASO performed by MC under direct supervision by one of the senior colleagues (this included 3 simple and 2 complex TGA). Since this time MC has performed a further 16 ASO independently. The overall number of ASO performed by the mentored surgeon over a 24 month period was 21 procedures.

Table3 shows the comparison between the patient’s characteristics and the intra and post operative results of the mentored (MC) and the mentoring (AP, AJP) surgeons. As expected the operative times (CPB and cross clamp time) were longer for the mentored compared with the mentoring surgeons but this did not translate into any significant differences in terms of mortality and early morbidity between the two group of surgeons. In fact hospital stay was noted to be significantly shorter for patients operated on by the mentored compared with the mentoring surgeons.

Table. 03

Comparison of early outcomes between mentored and mentoring surgeons

Mentored Surgeon Mentoring Surgeons P value
Number of ASO 21 117
Age (days) 10 (7-15) 11 (8-15) 0.6
Weight (kgs) 3.5 (3.2-3.7) 3.4 (3.1-3.7) 0.7
Unusual coronary anatomy 10 (47.6%) 24 (21.3%) 0.01
Complex TGA 5 (23.8%) 36 (30.7%) 0.5
In-hospital mortality 0 2 (1.7%) 0.5
Cross-clamp time (min) 91 (84-96) 60 (51-76) 0.0001
CPB time (min) 133 (120-163) 104 (78-127) 0.0001
Difficult weaning from CPB 2 (9.5%) 5 (4.2%) 0.3
Ventilation time (hours) 51 (48-72) 48 (30-72) 0.06
ICU stay (days) 4 (3-6) 4 (3-8) 0.5
Post-operative complications Major 2 (9.5%) 8 (6.8%) 0.3
Minor 0 13 (11.1%)
Hospital stay (days) 10 (8-15) 14.5 (12-20) 0.005

Note. ASO=arterial switch operation; CPB=cardiopulmonary bypass; TGA=transposition of the great arteries; ICU=intensive care unit.


What emerged from the BRI Inquiry 1,2 was a comprehensive report highlighting many of the deficiencies which led to poor surgical outcomes especially in complex congenital heart defects such as TGA. Some of the problems were related to individual failing in terms of lack of insight, leadership and teamwork, while others were due to inadequate facilities, the service offering paediatric open-heart surgery being split between two sites, and no dedicated paediatric intensive care beds. The current report highlights the significant progress that has been made in the last decade in Bristol, following rebuilding of the service.

Our results compare favorably both with national6 ,11 and international 1215 major series of ASO. The Central Cardiac Audit Database (CCAD) in collaboration with the Society for Cardiothoracic Surgery and the British Congenital Cardiac Association provides a profile of every congenital heart disease centre in the UK from 2000 onwards. 6,11 The 30 day survival after a simple or complex ASO was 96.9% and 92.8% respectively in the 16 different UK centers. In Bristol the 30 day survival after a simple or complex ASO was 100% and 95.2% respectively.

As expected the early outcomes of complex TGA were worse compared to simple TGA, and the most important cause of mortality was related to coronary complications, specifically intramural coronary anatomy, a known independent risk factor for early mortality16 and possibly a contraindication to the ASO. In our study the number of patients with an intramural course of the coronary arteries is too small to draw any significant conclusions, but we are evaluating the possibility of alternative surgical strategies.

At follow up there were 2 deaths in the first year with a subsequent zero mortality after 12 years. Mortality was related to major events that occurred immediately after the ASO. Cardiac related events at follow up were mainly associated with pulmonary stenosis (6.6%) requiring catheter stenting or balloon dilatation. Only one patient required surgical pulmonary artery reconstruction. Good left ventricular function and sinus rhythm were maintained in almost all patients after ASO confirming again the efficacy of this surgical procedure. Interestingly one patient with an uneventful ASO underwent heart transplantation for unexplained progressive dilated cardiomyopathy unrelated to coronary artery disease.

Our results with the ASO confirm previous reports documenting our radical improvement in surgical outcomes in neonates and infants. Aylen et al.17 reported trends in mortality of open cardiac surgery in children in Bristol and England since 1991 and observed that following the reconformation of the service mortality for open procedures in children aged under 1 has fallen markedly, from 29% (21% to 37%) before 1995 to 3% (1% to 6%) in the recent era. Similarly the operative mortality for the ASO dropped from 30% before 1995 to 1.4% afterwards, with no mortality for simple TGA undergoing ASO.

These improvements are the results of us implementing the recommendations of the Paediatric and Congenital Cardiac Services Review2 concerning the “patient journey”. It starts with the early antenatal identification of TGA which enables medical staff to make decisions, in discussion with the parents, about the management of the pregnancy and arrangements for delivery. The patient is then referred to a paediatric cardiologist who liases with the paediatric cardiac unit to discuss and plan the surgery. A series of protocols have been put in place and can be summarize as follows:

1. Retrieval of the patients by the PICU retrieval team which include both a senior and a junior doctor.

2. Immediate referral to the consultant cardiologist for echocardiological assessment.

3. Proceeding to balloon atrial septostomy in the majority of cases.

4. Daily echocardiographic assessment and discussion in the multidisciplinary meeting regarding the planning and timing of surgery.

All this demonstrates the fact that the establishment of a new service with dedicated specialists achieved immediate and complete changes within a very short time scale with significant benefits for our patients.

It has been recommended by the General Medical Council and the Royal Colleges of Surgeons of England that the individual surgeon’s performance should be monitored and methods for assessing their performance developed.18 For adult cardiac surgery, monitoring is already common.19 For congenital heart defects, such methods are not as well developed, and there are no guidelines regarding training newly qualified paediatric cardiac surgeons to complex operations such as ASO,20 although interesting work in this area has been done by de Leval and colleagues.21 Despite the continued need for obtaining new knowledge and learning new skills, the professional and public tolerance for a “learning curve” is much less than in previous decades. Still, professionals must learn to perform and develop independence and confidence. Mentorship from senior surgeons is therefore the key to success in this environment. Our results show no significant difference in outcomes during the training and introduction of a newly qualified paediatric cardiac surgeon to the unit. The success of this training program was achieved gradually through a sharing of responsibilities by both the mentor and the mentee. The junior surgeon had to develop a relationship of trust with the mentor in order to become a desirable candidate for potential mentorship. This was achieved by the mentee mirroring the skills of the senior surgeon and demonstrating great flexibility and availability in the learning process. The mentor had responsibilities as well, including a commitment to teach skills and transfer knowledge and responsibilities to the mentee. It was therefore the responsibility of both mentee and mentor to ensure a bidirectional exchange of benefit. This relationship required time, patience, dedication, and to some degree selflessness, but proved to be the best tool for mastering these complex surgical skills and maturing through the learning curve.


Since the “Bristol scandal” in the early nineties, early and long term mortality after the ASO at the BRHC has fallen markedly to below the national average. Improved quality of care together with a constant effort towards training and research may account for the fall in mortality through new technologies and improved perioperative and postoperative care.

This paper demonstrates that even in a unit which is under great scrutiny, not only can excellent results be achieved but good mentoring and training opportunities can be offered that help a new surgeon progress through the learning curve in a controlled and safe fashion, thus making the “Bristol scandal” a story with a happy ending.