Potential eligibility of congenital heart disease patients for subcutaneous implantable cardioverter-defibrillator based on surface electrocardiogram mapping

Abstract

Aims

The eligibility of complex congenital heart disease (C-CHD) patients for subcutaneous implantable cardioverter-defibrillator (S-ICD) has yet to be determined. The aim of this study was to determine in C-CHD patients: (i) the S-ICD eligibility, (ii) the most effective sensing vector, (iii) the impact of posture change on screening eligibility, and (iv) the impact of using two vs. six postures for screening. Adults with structurally normal hearts were used as controls.

Methods and results

The Boston Scientific ECG screening tool was used to determine eligibility for S-ICD in two and six different postures in 30 patients with C-CHD and 10 controls. Statistical significance was determined using Fisher's exact test. In total, 1440 bipolar vectors were collected. The mean age was 36.3 years, 57% subjects were men. Over all 86.7% of C-CHD patients and 100% controls (P > 0.05) met S-ICD eligibility. In controls, the primary vector (PV) was the most effective, and the alternate vector (AV) was least effective. In C-CHD patients, the AV was comparable to the PV. Posture change did not significantly affect S-ICD eligibility in C-CHD patients and controls (P > 0.05). Screening with six postures vs. two did not significantly affect S-ICD eligibility of C-CHD patients (83% vs. 87%, P > 0.05) or controls (90% vs. 100% P = >0.05).

Conclusion

No significant differences were observed between S-ICD eligibility in C-CHD patients and controls. The AV and PV are most suitable in C-CHD patients. No significant impact of postural change was observed for S-ICD eligibility between the two groups. No significant difference was observed in S-ICD eligibility when screening using two or six postures in both groups.

Introduction

The role of implantable cardioverter defibrillator (ICD) in primary and secondary prevention of sudden cardiac death is well established.^1–4 However, the conventional ICD requires placement of endovascular leads in the right side of the heart. Such leads offer stable electrogram sensing ability but the implant procedure to site the endovenous lead carries the risk of complications such as infection, pneumothorax, myocardial perforation, tamponade, and vessel trauma.⁵ The leads themselves are at risk of acute dislodgement as well as longer term complications, such as disruption of lead insulation and lead fracture.⁵ These issues are of greater importance in young patients who are likely to require multiple generator exchanges and lead replacements with the increased risk of complications associated with these procedures. Younger patients have a higher incidence of lead-related complications such as lead fracture as they are more active and traditionally have had smaller diameter leads implanted.^6,7 Failed leads may result in inappropriate shocks, and management includes lead extraction with it's attendant risks.⁶ These factors may militate against optimal ICD uptake and have acted as a stimulus for the development of subcutaneous ICD (S-ICD).⁸ Unlike conventional ICD, the S-ICD detects cardiac rhythm change by far-field sensing of cardiac electrical activity using three subcutaneous (body surface) sensing electrodes. The S-ICD sensing algorithm discriminates normal rhythm from arrhythmia through the analysis of this sensed subcutaneous electrocardiographic (ECG) morphology.^9,10 However, it is inevitable that the surface ECG morphology may vary due to anatomical differences between patients and within patients due to posture-related change in cardiac orientation relative to the fixed subcutaneous sensing electrodes.¹¹ This is likely to be particularly pertinent in the anatomically heterogeneous congenital heart disease (CHD) population.

Some subgroups with CHD are at an elevated risk of arrhythmia and sudden death as a consequence of cardiac structural abnormalities, pressure and volume overload, myocardial fibrosis, and myocardial scar related to palliative surgery.^12,13 However, the rate of ICD implant in this group is relatively low.¹⁴ This is partly due to poor risk stratification,¹⁵ and also because of young age (higher risk of endovenous lead complications) and implantation challenges due to naturally occurring or surgically created barriers to safe endovenous lead placement.⁶ Furthermore, in CHD subgroups with intra-cardiac shunts and single-ventricle anatomy, there is higher risk of thromboembolism in the presence of a transvenous lead.¹⁶ Conceptually the CHD patients may be a target population for S-ICD placement, but it is unclear whether the S-ICD sensing algorithm is suitable in the context of CHD anatomies, notwithstanding evidence for defibrillation efficacy of this technology in this context.⁸

The aim of this study was to determine in high-risk complex CHD patients: (i) suitability for S-ICD, (ii) the most suitable sensing vector, (iii) the impact of posture change on screening, and (iv) differences due to screening between use of two postures compared with six postures and using morphologically normal adults as a reference group.

Methods

This study was undertaken in the Wessex Cardiothoracic Centre, University Hospital, Southampton, which is a tertiary care centre for adult cardiology, paediatric cardiology, and adult CHDs.

Ethical consideration

The study received approval from an independent review board of the Southampton and South West Hampshire Research Ethics Committee B (REC 08/H0504/55) and the Research and Development Department of University Hospital, Southampton NHS Trust (UHS). All subjects included in the study gave full informed consent.

Study population

All the subjects were aged 18 years or over and had the ability to give informed consent. Forty subjects were recruited.

Group 1. Adults with morphologically normal heart on cardiac magnetic resonance imaging (N = 10). (reference group).

Group 2. Adult CHD patients (N = 30), including 10 patients with tetralogy of Fallot (TOF), 10 patients with transposition of great arteries (TGA), and 10 patients with Fontan circulation and single-ventricle physiology (SVP).

These three groups of CHD represent relatively common, complex patients with elevated risk of sudden death.¹⁷

Patients in an arrhythmia or a paced rhythm were excluded from the study.

Electrocardiographic data collection

Previous studies have validated body surface ECG as an adequate surrogate for subcutaneous ECGs.¹⁸ Therefore, the manufacturer (Boston Scientific) of the only commercially available S-ICD recommends pre-implant screening using a three-lead surface ECG in two postures (standing and supine). Using this pre-implant screening method, bipolar vectors were collected with a three-channel bipolar ECG at a sweep speed of 25 mm/s, using a sampling rate of 1 kHz, and an ECG gain between 5 and 20 mm/mV, for 10 s, in 6 postures (standing, sitting, supine, left lateral, right lateral, and prone). Prior skin preparation was carried out (alcohol wipe and shaving hair where necessary) to allow adequate adhesion of individual ECG skin electrodes and high-quality signal collection. Three bipolar electrodes (commonly known as LL, LA, RA) of the standard ECG machine (GE MAC 5500, USA) were used for data collection. The electrode LL was placed in the fifth intercostal space along the left mid-axillary line, electrode LA was placed 1 cm left lateral of the xiphoid and electrode RA was placed 14 cm superior to the LA electrode, 1 cm left lateral to the sternal margin (Figure 1). The bipolar vector lead I was derived from RA and LA, lead II from RA and LL, and lead III from LA and LL, representing surface ECG equivalent of Boston Scientific sense vectors (primary = lead III, secondary = lead II, alternate = lead I). A single investigator collected the data and two trained investigators, who were blinded to the patients' details, analysed all ECGs separately.

ECG analysis

The manufacturer of the currently available S-ICD (Boston Scientific) has developed an ECG morphology-based pre-implant screening tool to identify patients with acceptable sensing characteristics prior to the implant of S-ICD.¹⁹ This is a template (printed chart) containing six coloured profiles of varying sizes, simulating the automatic gain adjustment function of S-ICD. This template has a horizontal line passing through all the colour profiles for adjustment of the isoelectric baseline. Each colour profile has an identical window above and below the baseline to account for positive and negative amplitude of R-wave and T-wave contours. Each window is subdivided by a dotted line and the peak of R-wave has to lie within this sub-window for one of the six profiles to be appropriate for sensing. Additionally, the trailing T-wave has to be contained within the same colour profile as the R-wave for the vector to be appropriate for sensing (Figure 2).¹⁹ This screening tool was used to evaluate each sense vector. QRS complexes, with minimum noise, were analysed for each vector. For biphasic signals, the larger peak was used to determine the appropriate colour map. The left edge of the selected coloured map was aligned with the onset of the QRS complex. If, when printed at the maximum 20 mm/mV gain, the QRS peak did not reach the minimum boundary (dotted line) of the smallest coloured profile, the vector was considered unacceptable. If the entire QRS complex and trailing T-wave were contained within the coloured profile, the vector/posture combination was considered suitable. If any portion of the QRS complex or trailing T-wave extended outside of the coloured profile, the sense vector was considered unacceptable. All vectors were examined individually.

The data were analysed with the screening template on the basis of six postures (standing, sitting, supine, left lateral, right lateral, and prone) as well as the currently used conventional pre-implant analysis of two postures (standing and supine).

Criteria for subject and vector suitability

A patient was considered a candidate for S-ICD implant if at least one and the same sense vector was acceptable for all tested postures.¹⁹ Similarly, any given vector was considered suitable if it satisfied the screening tool in all tested postures¹⁹ (e.g. a patient/subject who had the primary vector suitable, then that subject had to have the primary vector suitable in all the tested postures, if the patient had the primary vector unsuitable in any of the tested posture but had secondary vector suitable in the given posture despite that the subject/patient was declared unsuitable for the reason that the current generation of S-ICD is limited in its ability to automatically switch between sensing vectors).

Statistical analysis

Statistical analyses were performed using the SPSS 20.0 software package (IBM SPSS Limited). Continuous variables are expressed as mean ± 1 SD. The proportion of patients fulfilling the ECG criteria is presented as percentages with 95% confidence interval (95% CI) using the Wilson procedure without a correction for continuity as described by Robert Newcombe.²⁰ Fisher's exact test was used to determine significant differences. The suitability of CHD groups was compared against normal control, and the suitability of leads was compared against lead III (primary vector). The suitability of postures was compared against supine posture. A P < 0.05 was considered significant.

Results

The mean age of the subjects was 36.3 ± 14.4 years and 57% were men. A total of 1440 vectors were obtained from 40 subjects in six postures, through three electrodes, in at least two gain settings. Using the pre-implant screening mapping system 750/1440, 52% (95% CI: 49–54) vectors were suitable for S-ICD sensing. For the purpose of analysis, a vector suitable at one or more gain setting was counted as +1 and a vector not suitable for sensing at any gain setting counted as −1. Thus, the final number of vectors for the analysis was 720, and 494/720, 69% (95% CI: 65–72) vectors were suitable either at one or more gain settings.

Eligibility of congenital heart disease patients for subcutaneous implantable cardioverter-defibrillator in comparison to normal controls

The proportion of CHD patients meeting the S-ICD screening criteria was compared with normal controls.

On the basis of conventional two-posture screening (standing and supine), 100% (95% CI: 72–100) subjects with structurally normal heart and 87% (95% CI: 70–95) CHD patients [TOF 80% (95% CI: 49–98), TGA 100% (95 CI: 72–100), and SVP 80% (95% CI: 49–98)] met the S-ICD pre-implant screening criteria (all P = > 0.05 by comparison against control) (Table 1).

On the basis of six postures, 90% (95% CI: 60–98) subjects with structurally normal heart, and 83% (95% CI: 66–93) of CHD patients [SVP 80% (95% CI: 49–98), TGA 90% (95% CI: 60–98), and TOF 80% (95% CI: 49–98) (all P = > 0.05 by comparison against control)] met pre-implant S-ICD ECG screening criteria (Table 1).

Effect of postures on vector suitability

Figures 3 and 4 and Table 2 show the impact of posture on vector suitability. The variation of vector suitability in different postures was statistically insignificant (P > 0.05) both in individual with normal cardiac morphology and CHD.

Screening with two and six postures

The differences in eligibility due to screening with two postures and six postures in normal subjects (100% vs. 90%) and patients with CHD (87% vs. 83%) were statistically insignificant.

Differences in leads

Tables 1 and 2 show the suitability of leads in the four groups, in two and six postures. In normal subjects lead III (primary vector) was the most suitable vector (90%), followed by lead II (secondary vector) (80%), while lead I (alternate vector) was least suitable (30%). However, in CHD patients lead I (alternate vector) suitability (67%) was comparable to lead III (73%) and better than lead II (46%). Overall the suitability of lead III was statistically superior to leads I and II (P = < 0.05). Table 3 shows the reason for vectors failure.

Suitable number of vectors

Table 1 shows the number of suitable vectors in normal subjects and subjects with CHD in two and six postures screening. The suitability of one, two, and three vectors was fairly uniform across all the groups, without any statistically significant difference.

Discussion

This study has demonstrated no significant differences in eligibility of complex CHD patients (TOF, TGA, SVP) and subjects with structurally normal hearts for commercially available S-ICD using the pre-implant ECG screening criteria. However, we have demonstrated that while CHD patients and normal controls met the S-ICD implant criteria using the primary vector more frequently, the alternate vector is more suitable in CHD patients. This study has also demonstrated the impact of body posture on sensing vector choice (Figure 3 and Table 2), and consequently, potential lead location in the thoracic subcutaneous tissues. The primary vector lead III was suitable in most postures in most cases, including controls as well as all groups of CHDs; this may be due to horizontal orientation of this lead in comparison to vertical and diagonal orientation of leads I and II. Lead I suitability varied with posture in normal subjects. However, lead I showed less variation with posture change in individuals with CHD (Figure 3), and specifically in patients with TOF, we speculate that right ventricular hypertrophy that occurs in patients with CHD and specifically TOF may have this effect on sensing vector effectiveness. Lead II showed variation with postural change in all CHD groups and specifically in individuals with SVP and was least suitable in this group (Figure 3). The screening with six (standing, sitting, supine, left lateral, right lateral, and supine) and two postures (standing and supine) showed a trend towards higher suitability at two postures. However, the difference was not statistically significant. Adding an additional four postures excluded 2 of 40 patients (5%) due to T-wave enlargement (oversensing) in the left lateral posture in the affected subjects. These two subjects would have otherwise satisfied the S-ICD implant criteria. This has potentially important implications for T-wave oversensing susceptibility that could be reduced by ECG screening with six postures.

The current body of published literature in relation to experience of S-ICD in CHD is limited.^21,22 The recently published early results from the EFFORTLESS S-ICD registry had only 7% (33 of 472) patients with CHD. However, the detailed description of the underlying CHD anatomy, the pre-implant screening results, fall out rate, reason for S-ICD implant, and the performance of S-ICD have not been described separately.²³

Table 3 shows the reasons for vectors failure. The screening tool is used as a proxy to identify QRS-T complexes that are likely to satisfy the S-ICD sensing algorithm and avoid inappropriate sensing performance. An R/T ratio <3 in the lead with the largest T-wave on the standard surface 12-lead ECG has been identified to be a strong predictor (odds ratio 14.6) of failed QRS-T morphology screening for the S-ICD.²⁴ We have found that an R/T ratio alone is less predictable of vector suitability as despite suitable R/T ratio the overall vector may not be suitable for sensing due to very large or small amplitude of both or individual R-wave and T-wave. The vector suitability also depends on the QRS duration and QT interval, as prolongation of these intervals makes the T-wave unsuitable. Additionally a borderline R/T ratio that satisfies the screening tool may increase the risk of future inappropriate S-ICD shocks due to T-wave oversensing in comparison to clearly suitable R/T ratio; in such patients screening with exercise test may be helpful.

Suitability of more than one vector would make more stable sensing possible; however, the current generation of S-ICD is limited in its ability of automatic mode switching between sensing vectors, and this has to be done manually with the device programmer, thus in current settings the suitability of multiple vectors have limited role.

Our study has several limitations. First, the sample size is small; it is possible that larger numbers may have revealed the smaller differences in eligibility of the groups studied to be significant. However, this study was designed to demonstrate any major differences between normal subjects and patients with complex CHDs. Moreover the number of complex CHD patients attached to any single centre is small and difficult to recruit; therefore 10 near age- and sex-matched subjects were recruited from normal control, TOF, TGA, and SVP to reduce the compounding factors. Moreover, prior to the result of this study power calculation would not have been possible. In addition, all data were collected by a single investigator to reduce variation, and furthermore, the sample size for this study was selected to mimic preclinical drug safety studies.²⁵ Second, since ECGs were collected from individuals in sinus rhythm at rest, there is possibility of variation in the morphologies of ECGs during exercise and arrhythmia. However, in this study the Boston Scientific S-ICD pre-implant screening method was followed, which recommends collection and analysis of resting surface ECGs and more recently screening is also performed on ECGs acquired during exercise.¹⁹ Third, the pre-implant screening process also assumes that vectors suitable at pre-implant screening fulfill criteria for the sensing algorithm of the S-ICD when implanted and forms the basis of implant decision-making. The defibrillation ability of S-ICD is beyond the scope of this study.

Conclusion

Using current pre-implant screening criteria, no statistically significant differences were observed between the proportion of CHD patients meeting S-ICD screening criteria and normal controls. No statistically significant impact of postural change was observed on eligibility of normal subjects and patients with CHD. Lead III (primary vector) met the screening criteria more frequently in CHD patients and normal controls. Lead I (alternate vector) was least suitable in subjects with structurally normal heart; however, this was more often suitable in CHD patients. Screening at two and six postures had no statistically significant effect on suitability of either normal subjects or patients with CHD using the current ‘conventional’ screening approach, but we speculate that screening with six postures could reduce the problem of T-wave oversensing. This hypothesis requires further evaluation but is an important observation given the impact of T-wave oversensing on inappropriate shock therapies in S-ICD recipients.

Conflicts of interest: J.M. and P.R. have received Honoraria and research grants from Medtronic, St Jude, Sorin, and Boston Scientific. N.C. has received unrestricted research grant from St Jude Medical, Haemonetics, and Medtronic. N.C. has also received honoraria/speech fees from St Jude Medical, Haemonetics, Abott Vascular, Boston Scientific, and Heart Flow. D.W. and M.Z. are supported by educational grants from Medtronic.

Funding

This study was partially funded by an unrestricted research grant from Medtronic UK and Cameron Health/Boston Scientific. PRIME ECG® vests were provided by an unrestricted grant from Heartscape Technologies/Verathon.

References