OUP user menu

Cardiac defibrillation therapy for at risk patients with systemic right ventricular dysfunction secondary to atrial redirection surgery for dextro-transposition of the great arteries

Kevin A. Michael, Gruschen R. Veldtman, John R. Paisey, Arthur M. Yue, Stephen Robinson, Stuart Allen, Nadia S. Sunni, Chris Kiesewetter, Tony Salmon, Paul R. Roberts, John M. Morgan
DOI: http://dx.doi.org/10.1093/europace/eum001 281-284 First published online: 23 March 2007


Aim To review techniques of implantable cardioverter-defibrillators (ICD) in patients after Mustard surgery for arterial transposition.

Methods and results Retrospective analysis of all Mustard patients receiving ICDs at our institution. Five patients (median age 24 years, range 19–35, 3 male) with systemic right ventricular dysfunction (sRV) dysfunction and New York Heart Association (NYHA) II and III, received ICDs. Implantation was performed transvenously in three patients, epicardial patches and subcutaneous arrays at surgery in two patients. Two patients required lead extraction and baffle stent angioplasty before ICD implantation. Defibrillation vectors incorporating the anterior sRV mass [i.e., sub-pulmonary left ventricle (pLV) to generator can, and between epicardial defibrillator patches], consistently achieved a minimum 10 joule(J) safety margin during defibrillation threshold (DFT) testing. Subcutaneous arrays and endocardial vectors that included a superior vena cava (SVC) electrode were less effective. One patient developed pulmonary oedema post-procedure. At a median 20 months, all patients were alive and in NYHA class II. Follow-up over 24 months documented multiple non-sustained ventricular tachycardia (VT) in the group and one patient had recurrent VT with aborted device therapy.

Conclusion Defibrillator implantation in Mustard patients is challenging. Sub-optimal defibrillation should be anticipated and can be overcome using vectors which integrate the RV mass and high-energy devices. A staged procedure involving pre-implant interventions or separate DFT tests, where indicated, may be better tolerated by patients.

  • Dextro-transposition of the great arteries
  • Mustard operation
  • Sudden cardiac death
  • Implantable cardioverter- defibrillator


Atrial redirection surgery, brought into common use by Mustard's modification of the procedure, transformed the natural history of dextro-transposition of the great arteries (d-TGA) from a 90% 1 year mortality to a 90% 1 year survival.1 Although rendering the infant acyanotic with excellent functional capacity, the morphological right ventricle (sRV) is required to support the systemic circulation following the Mustard procedure. The sRV faces substantial pressure overload resulting in a high incidence of late sRV failure.2

Mustard patients have been demonstrated to be at increased risk of sudden cardiac death (SCD) proportional to the degree of sRV dysfunction as well the presence of atrial arrhythmias.3 In other populations recognized to be at high risk of SCD, it is well recognized that ICD therapy is a highly effective and arguably the only definitive treatment modality.49 Given the particular anatomy after a Mustard procedure, problems with transvenous access to the pLV and altered vectors of defibrillation, it cannot be assumed that the general efficacy of ICDs will be reproduced in these individuals. In order to describe the technical considerations involved, we review a single centre experience of ICD therapy in five consecutive patients with a Mustard procedure for d-TGA considered to be at high risk of SCD.


A retrospective review of hospital records was conducted to evaluate technical considerations, implant details, and status at follow-up on five patients with d-TGA and a Mustard procedure receiving an ICD ± concomitant cardiac resynchronization therapy (CRT). Functional status, arrhythmia history, and echocardiographic data before implant were reviewed. An individualized approach to implantation is practised in our institution, taking into consideration existing transvenous electrodes and post-surgical cardiac anatomy. Following implant, patients were followed up at 4–6 weeks and then 6-monthly intervals. Data at follow-up including functional assessment, device interrogation, ECG, and echocardiograms were reviewed.


The baseline characteristics of the 5 patients (4 male, age 18–35 years) studied are shown in Table 1. All had significantly impaired sRV function [mean sRV ejection fraction (EF) = 30%] and impaired functional class (NYHA class II and III). Two patients had sustained VT, three non-sustained VT, and four atrial arrhythmias. Three patients had previously implanted pacemakers. Two patients were implanted with combination CRT/ICD devices. Pre-implant electrophysiological studies were not performed.

View this table:
Table 1

Salient clinical characteristics at the time of implantation

Patient 1Patient 2Patient 3Patient 4Patient 5
Age (years) at implantation2422351825
Documented arrhythmiasSinus bradycardia, atrial flutter, NSVTSinus bradycardia, VT, NSVTSinus bradycardia, atrial flutter, NSVTSVT, NSVTSinus bradycardia, SVT, NSVT, VT
RVEF (%)2525303535
Pre-existing deviceYesYesYesNoYes
Weight (kg)110.2106.17767.454
  • NSVT, non-sustained ventricular tachycardia; SVT, supra-ventricular tachycardia; RVEF, right ventricular ejection fraction.

Implant considerations

Implanting a transvenous electrode into the pLV requires that lead navigate the surgical baffle. Baffle stenosis impeding flow from the superior vena cava (SVC) to the pLV is common post-Mustard modification.1 Defibrillator electrode placement will potentially further impede venous drainage and pLV filling. A strategic decision was therefore made not to implant multiple leads via anatomically narrow baffles and to treat any baffle stenosis prior to the implant. An exclusively transthoracic defibrillation strategy was thus employed in the first patient using a high-energy defibrillator (maximum output of 41 J) with subcutaneous, single finger arrays (Model 6996: 25 cm coil, 500 cm2 surface; Medtronic, MN, USA). Defibrillation testing was not performed at the end of the device implant because of haemodynamic instability during a lengthy procedure but at a separate session. The vulnerable sRV may not have tolerated induced ventricular fibrillation (VF). Unfortunately, when tested, defibrillation was inconsistent. In view of our subsequent success with endocardial defibrillation, we plan to include a transvenously implanted defibrillation lead in pLV, in the shock configuration.

Of the four patients with pre-existing pacemakers, two were demonstrated to have functional baffle stenosis and therefore underwent lead extraction and stent angioplasty prior to transvenous implant of a defibrillator lead to the pLV. Those patients who had lead extractions and the patient with no pre-existing pacemaker underwent standard transvenous implants with endocardial defibrillation coils placed via the subclavian vein into pLV. Two patients with existing pacemakers and no evidence of baffle stenosis at cardiac catheterization had their existing pace/sense electrodes preserved; one had epicardial defibrillation patch electrodes and one had subcutaneous arrays implanted at the time of trans-septal, epicardial CRT.

Defibrillation circuits and testing

The patient with subcutaneous arrays had the electrodes placed in an anteroposterior configuration with a left subpectoral active can implantation (Table 2). Ventricular fibrillation was induced via the endocardial pace/sense lead and therapies were delivered between the anterior and posterior arrays. A series of 6 inductions were performed with therapies of 25–41 J, but none was successful in terminating VF and external rescue shocks were required.

View this table:
Table 2

Characteristics of defibrillation tests

Patient 1Patient 2Patient 3Patient 4Patient 5
Defibrillation vectorAnterior and posterior subcutaneous arraysAnterolateral and inferior epicardial patchesVentricular coil in pLV to can (SVC excluded)Can to coil in pLV (SVC excluded)Can to coil in pLV (No SVC coil)
Number of shocks62432
Total energy delivered during defibrillation test262 J50 J100 J75 J50 J
Successful 25 J shockNoYesYesYesYes
Average Impedance (ohm)873.563.066.349.371.0

One patient underwent epicardial patch electrode placement in the anterolateral and inferior positions over the systemic ventricle. This strategy was successful in defibrillating VF at 25 J on the first induction (a 10 J safety margin).

The remaining three patients underwent standard transvenous system implants. An active can was placed in the left subpectoral region and a defibrillation coil advanced under fluoroscopy to the pLV via the subclavian vein. The first two transvenous implants received dual coil leads but both had failed initial defibrillation, requiring exclusion of the SVC coil from the circuit. In one of these patients, a pLV to can vector provided effective defibrillation at 25 J, and in the other it was also necessary to reverse the polarity of the circuit to achieve this safety margin. This patient underwent four VF inductions and subsequently developed pulmonary oedema during the procedure.


The median follow-up in all 5 patients was 20 months (range 15–24). A total of 4083 non-sustained VTs have been documented in the cohort over this period. One patient has had multiple episodes of VT and pre-syncope which spontaneously resolved aborting a device discharge.


Mustard patients are at an increased risk of SCD defined by sRV dysfunction, heart failure functional status, non-sustained VT, and atrial arrhythmias.3 Only one patient in the cohort had a preceding out of hospital arrest, but all patients had documented multiple ventricular ectopics on 24-h Holter anaylsis. None of the patients received formal electrophysiological studies, as we felt that a negative study should not disqualify them. Our patients were young (<40 years of age) and lived productive lives up to the point of referral. The decision to implant these patients with ICDs without further investigation to define individual risk was in our estimate pragmatic and cost effective. During the follow-up (24 months), although numerous episodes of non-sustained VT have been documented, one of the patients has had recurrent VT. Considering that the time to the first appropriate therapy in other patient groups with ICDs often exceeds 2 years, this may justify our decision not to undertake more extensive pre-procedural risk stratification.10

It has also not previously been demonstrated that the gold standard treatment for prevention of SCD, namely ICDs, is effective in this population, and given the variation from normal anatomy, efficacy should not be assumed. Appropriate vectors of defibrillation, as determined by electrode placement, are the key determinants of defibrillation success.112 In the case of a pLV, the vectors developed from experience in acquired heart disease, have predictable shortcomings. The SVC cavity to pLV vector in particular appears to exclude much of the sRV and septum (Figure 1), while severe chamber dilatation, common in Mustard patients, is recognized from other populations to be associated with less effective defibrillation.13

Figure 1

A dual coil defibrillator lead implanted into the pLV tends to be located more superiorly than when it is in an apical position in the right ventricle. The result is that the proximal electrode rests within the SVC and left subclavian vein. The defibrillation vector between the SVC and pLV thus excludes the major myocardial mass, reducing defibrillation efficacy.

The role of polarity reversal is less well established. Data from randomized trials provide some evidence but no rational explanation, that by inverting vectors, efficacy may be improved in a patient.1415 However, with a limited number of inductions in any given individual, apparent differences in efficacy may be no more than a reflection of the probabilistic nature of defibrillation success, and we need to be cautious in making finite decisions as regards what will and will not work in individual patients.

The single patient experience with subcutaneous defibrillation was plainly suboptimal with no successful defibrillation achieved. This may be a reflection of the Mustard anatomy or other independent factors that potentially may affect defibrillation efficacy: this patient had a BMI of 33 and gross cardiomegaly.16

Cardiac resynchronization in this group was performed as an adjunct to ICD implantation and presented its own challenges, eventually requiring a hybrid approach with endocardial/epicardial systems.

The central issue explored in this report is how to manage this population at increased risk of SCD, with particular challenges to effective defibrillator implantation. Our experience demonstrates that transvenous implantation is feasible, but may require interventions such as lead extraction and angioplasty of the atrial baffle to avoid venous pathway obstruction. The optimal defibrillation vector with a transvenous implant appears to be pLV to active can and as such, single coil leads are preferred. Our experience with subcutaneous arrays alone was unsatisfactory. Efficacy may be improved by selecting patients with a smaller body habitus and by incorporating intra-thoracic electrodes. Epicardial patches require an open chest procedure, but may be placed at the time of concomitant epicardial CRT if indicated in the individual patient.

Our findings mandate the use of high-output devices, regardless of the configuration, in anticipation of compromised defibrillation efficacy in similar patients.

The patient who underwent a prolonged procedure prior to DFT testing resulting in post-procedural pulmonary oedema, emphasized the inability of the compromised myocardium to sustain stresses. We, therefore, now undertake a staged approach to the ICD implant, deferring defibrillation testing to another session if the implant procedure was prolonged or complicated.


This is a descriptive report on a small cohort and as such does not provide sufficient quantitative data. Given the low likelihood of a randomized trial, multi-centre, pooled, observational data are necessary to obtain a greater understanding of device therapy in this population.


Defibrillator implantation in Mustard patients is feasible though challenging. Clinicians planning to implant such patients must be prepared to optimize the systemic access past the baffle prior to implant and then have strategies in place to tackle high-defibrillation thresholds and the decompensation that may result from multiple VF inductions. Endocardial pLV to active can defibrillation and epicardial patches are effective defibrillation strategies in this group.

Conflict of interest: P.R.R. and J.M.M. are currently conducting research sponsored by Medtronic.


View Abstract