Defibrillation Safety Margin Testing in Patients With Congenital Heart Disease: Results From the NCDR

Congenital and Pediatric Cardiac EP Feature

Defibrillation Safety Margin Testing in Patients With Congenital Heart Disease: Results From the NCDR.Prutkin JM, Wang Y, Escudero CA, Stephenson EA, Minges KE, Curtis JP, Hsu JC.JACC Clin Electrophysiol. 2021 Apr 22:S2405-500X(21)00213-9. doi: 10.1016/j.jacep.2021.02.019. Online ahead of print.PMID: 33933411

Take Home Points:

1. Defibrillation Safety Margin (DSM) Testing) at time of initial implant and generator change has decreased from 2010 to 2016 (64% to 31.4%)
2. DSM testing did not confer additional harm and was not associated with in-hospital complications or death
3. Inadequate DSM testing occurred in 13.7% and was not associated with in-hospital complications or death
4. Factors associated with DSM testing included race, NHYA functional class, atrial fibrillation, ejection fraction, primary prevention device, generator change, physician training, and US geographic region
5. Factors associated with inadequate DSM testing included race, body surface area, NHYA functional class, QRS duration, systolic blood pressure, ICD type, physician training, and US geographic region

 Commentary by Dr. Akash Patel (San Francisco, CA, USA) Congenital and Pediatric Cardiac EP section editor:  

Defibrillation threshold testing (DFT) has been routinely done at the time of implantable cardioverter defibrillator placement or replacement. Over time this practice has changed with ~ 50% of adult cases undergoing DFT testing. This has been due to the advent of higher voltage devices and adult studies showing the safety and efficacy of omitting testing. The impact of this change in device testing practices on patients with congenital heart disease is unclear.   In addition, traditional DFT testing to avoid multiple inductions and shocks has shifted to the use of defibrillation safety margin (DSM) testing with at least a 10J safety margin below the maximal output of the device. This study examined the prevalence of DSM testing at the time of ICD insertion, predictors of testing, factors associated with an inadequate DSM testing, and the rate of in-hospital adverse events associated with DSM testing in patients with CHD.

This was a retrospective study of the National Cardiovascular Data Registry (NCDR) ICD Registry between April 1, 2010, and March 31, 2016.  All congenital heart disease patients (pediatric and adults) were included. A total of 7024 patients were analyzed. The vast majority had a transvenous system with only 30 (0.4%) receiving an epicardial non-transvenous ICD system. DSM testing was performed in 3,654 patients (52%) and not performed in 3,370 patients (48%).

The demographics of the total cohort included atrial septal defects (2933, 41.8%), ventricular septal defects (2119, 31.3%), tetralogy of Fallot (1697, 24.2%), transposition of the great vessels (866, 12.3%), Ebstein anomaly (177, 2.5%), and common ventricle (49, 0.7%).  The median age was 53.1±18.1 years. Non-ischemic dilated cardiomyopathy was noted in 45.4% and heart failure was noted in 67.7%. Atrial fibrillation/futter was noted in 49.3% with 39.7% on warfarin.  Amiodarone was used in 13.4% and sotalol was used in 7.0% which are known to impact defibrillation thresholds.

The ICD device types included dual chamber (43.7%), CRT-D (37.4%), and single chamber (18.8%). ICD indications included primary prevention in 62.8% and secondary prevention in 37.2%.

The prevalence of DSM testing overall was 48%. There was a decrease in DSM testing from 64% in the 2^ndquarter of 2010 to 31.4% in the 1^st quarter of 2016. There was no difference in those patients who received devices for primary vs secondary prevention and initial implant vs at generator change.  See figures below.

Clinical factors associated with DSM testing included younger age (51.4 ± 18.0 vs 54.9 ±18.0 years), absence of heart failure (36.7% vs 27.5%), less severe NHYA functional class, higher LV ejection fraction  (38.4 ± 15.3% vs 35.8 ± 15.4%) absence of non-ischemic dilated cardiomyopathy (57.8% vs 51.1%), shorter non-paced QRS duration (136.6 ±35.8 vs 138.8 ±35.9 msec ), absence of atrial fibrillation/flutter (57.4% vs 43.4%), lower warfarin use (33.2% vs 46.7%), lower creatinine (1.08 ± 0.89 vs 1.13 ± 0.72), lower use of amiodarone (12,2% vs 14.8%) and higher use of sotalol (7.7% vs 6.3%).  Type of congenital heart disease was also associated with DSM testing. There were significantly less patients with atrial septal defects (40.2% vs 43.5) and more tetralogy of Fallot patients (26.2% vs 21.9%) in the testing group.  Device factors associated with DSM testing included type of device and ICD indication (primary prevention: 59.6 vs 66.3% and secondary prevention: 40.4% vs 33.7%). Additional factors associated with DSM testing included physician training (EP board certified, 74.5% vs 69.4%), insurance payer (private insurance, 44.4% vs 37.1%), and hospital owner (private, 68.9% vs 63.6%).   See tables below.

On multivariate analysis, clinical factors associated with lower likelihood of DSM testing included race (other), NHYA classification (III and IV), presence of atrial fibrillation, lower ejection fraction, and warfarin use. Device factors associated with lower likelihood of DSM testing included generator change and primary prevention ICD indication.  Additional factors associated with lower likelihood of DSM testing included procedure performed in Northeast US geographic region and physician training (EP fellowship only or surgery boards).  See table below.

On multivariate analysis for complex congenital heart disease, factors associated with a lower likelihood of DSM testing included atrial fibrillation, heart failure, warfarin use, generator change, primary prevention indication, and performed in Northeast US geographic region. Factors associated with a higher likelihood of DSM testing included body surface area. See table below.

For those who underwent DSM testing, the mean lowest successful energy tested was 20.7 ± 7.3 J.   An inadequate DSM was defined as a <10 J difference between the lowest successful energy tested during DSM testing and the maximum output of device. Of the 3,623 with DSM testing data, the prevalence of inadequate DSM was 13.8%. The mean lowest energy tested during DSM testing in the inadequate DSM testing group was 31.1 ± 4.0 J.

Clinical factors associated with inadequate DSM testing included younger age (49.8 ± 18.4 vs 51.6 ± 18.0 years), more severe NHYA functional class, longer non-paced QRS duration (140.7 ± 38.2 vs 136.0 ± 35.3 msec), and lower systolic blood pressure (122.6 ± 19.5 vs 124.6.0 ± 20.3 mmHg).  Type of congenital heart disease was also associated with inadequate DSM testing. There were significantly more patients with common ventricle (1.6% vs 0.7%) and Ebstein anomaly (4.6% vs 2.2%) in the inadequate DSM testing group.  Device factors associated with inadequate DSM testing included type of device (more single chamber devices, 23.6% vs 18.5%), higher lowest successful energy tested (31.1 ± 4.0J vs 18.8 ± 4.5J) and physician training (EP board certified, 73.1% vs 74.9%). See tables below.

On multivariate analysis, clinical factors associated with higher likelihood of inadequate DSM testing included race (black and non-Hispanic), NHYA classification (III and IV), and longer QRS duration. Device factors associated with higher likelihood of inadequate DSM testing included single chamber device.  Additional factors associated with higher likelihood of inadequate DSM testing included procedure performed in the Northeast and West US geographic region and physician training northeast US geographic region for location of procedure and physician training (pediatric cardiology boards).  See table below.

On multivariate analysis for complex congenital heart disease, factors associated with a lower likelihood of inadequate DSM testing included single and dual chamber devices. Factors associated with a higher likelihood of inadequate DSM testing included physician training (pediatric cardiology boards). See table below.

The overall adverse events included any complications in 2.3%, in-hospital mortality in 0.41%, and DSM-related complications in 1.2%.  DSM-related complications were defined as cardiac arrest, cardiac perforation, lead dislodgement, myocardial infarction, pericardial tamponade, peripheral embolus, transient ischemic attack or cerebral vascular accident, or urgent cardiac surgery.  There was no difference in complications or mortality between the DSM testing and no DSM testing groups on univariate and multivariate analysis.  Those with DSM testing had a shorter length of stay (LOS) (3.23 vs. 3.85 days) and a less prolonged post-procedural LOS > 3 days (8.1% vs 11.4%).   Of note, there was no differences in outcomes or post-procedural length of stay based on inadequate DSM testing. See table below.

Adverse event	Total (n=7024)	No DSM Testing (n = 3370)	DSM testing (n= 3654)	P
Any complication	162 (2.3%)	78 (2.3%)	84 (2.3%)	0.96
In-hospital mortality	29 (0.41%)	17 (0.50%)	12 (0.33%)	0.25
Any complication or death	173 (2.5%)	85 (2.5%)	88 (2.4%)	0.76
DSM-related complication*	82 (1.2%)	38 (1.1%)	44 (1.2%)	0.77
Length of stay days: Mean (SD)	3.53 (6.92)	3.85 (7.72)	3.23 (6.09)	0.0002
Post procedure length of stay days: Mean (SD)	1.63 (4.24)	1.78 (4.22)	1.49 (4.25)	0.0055
Prolonged post procedure length of stay > 3days	679 (9.7%)	384 (11.4%)	295 (8.1%)	<0.0001
Specific associated complication
Cardiac arrest	18 (0.3%)	11 (0.3%)	7 (0.2%)	0.26
Cardiac perforation	2 (0.03%)	2 (0.06%)	1 (0.03%)	0.14
Conduction block	1 (0.01%)	1 (0.03%)	0 (0%)	0.30
Coronary venous dissection	3 (0.04%)	2 (0.06%)	1 (0.03%)	0.52
Implant site hematoma	29 (0.41%)	16 90.47%)	13 (0.36%)	0.44
Hemothorax	4 (0.06%)	2 (0.06%)	2 (0.05%)	0.94
Infection requiring antibiotics	10 (0.14%)	3 (0.09%)	7 (0.19%)	0.25
Lead dislodgement	54 (0.77%)	21 (0.62%)	33 (0.9%)	0.18
Pericardial tamponade	2 (0.03%)	2 (0.06%)	0 (0)	0.14
Peripheral embolus	3 (0.04%)	2 (0.06%)	1 (0.03%)	0.52
Set screw problem	3 (0.04%)	0 (0)	3 (0.08%)	0.10
Pneumothorax	24 (0.34%)	13 (0.4%)	11 (0.3%)	0.54
Transient ischemic attack or stroke	3 (0.04%)	1 (0.03%)	2 (0.05%)	0.61
Urgent cardiac surgery	6 (0.09%)	2 (0.06%)	4 (0.11)	0.47
Venous obstruction	4 (0.06%)	1 (0.03%)	3 (0.08%)	0.36

Overall, this study demonstrated a decline in the use of DSM testing during ICD placement or generator changes to 31.4% in patients with congenital heart disease. This occurred overtime in conjunction with adult data and guidelines to support the potential role to omit testing. Of note, was the limited clinical data on DSM testing in congenital heart disease patients over the study time frame.

The study also demonstrated that the risks for adverse events with DSM testing were low and that DSM testing did not confer additional harm compared to omitting DSM testing. This finding is important, especially considering the higher rates of inadequate DSM testing in congenital heart disease patients (13.8%) compared to adults with transvenous ICDs (9.4%) and adults with subcutaneous ICDs (9.4%).

The study also demonstrated clinical, device, and provider factors that played a role in those who omitted or underwent testing and those who had inadequate DSM testing. DSM testing was more likely to be omitted in patients who underwent primary prevention ICD placement or were deemed higher risk due to severity of heart failure or associated comorbidities – atrial fibrillation and warfarin.  This avoidance in testing likely represented perceived concerns with procedural complications in DSM testing which was not supported by the findings of this study. Inadequate DSM was more likely to occur in patient who were black and non-Hispanic, had more severe electromechanical disease (heart failure or QRS prolongation) or received a single chamber device.  Social determinants of health including race have been shown to impact use and counseling for ICDs and need further exploration in congenital heart disease patients.   Device selection also played a role, but multiple implant variables were not accounted for in this study such as device location and maximal programmable energy delivered by devices which may have changed from 2010 to 2016 based on manufacturer.

This study does have some limitations.  First, the cohort was older and was not stratified by age (i.e., adult vs pediatric) which may play a role in DSM use, testing outcomes, complications, and length of stay.  Second, there was no mention of single vs dual coil leads, shock vector orientation, or location of the pulse generator (right vs left chest, abdominal, subpectoral vs prepectoral) which may play a role in the lowest energy used for successful DSM testing.  Third, the granularity and complexity of congenital heart disease, cardiac position (levocardia vs dextrocardia), associated cardiac surgery, and ventricular mass (i.e., associated hypertrophic cardiomyopathy) are unknown which would impact the use of DSM testing, results of DSM testing, and procedural complications which are solely defined as in-hospital. Fourth, it is unknown how many attempts for DSM testing were done and if the system was revised due to inadequate DSM testing.  Finally, this study was not designed to assess the most important question of whether omitting DSM testing is safe and non-inferior to DSM testing regarding clinical outcomes (i.e., appropriate and successful use of ICD therapy) and mortality.

Further clinical follow-up and studies are needed to understand the long-term clinical impact of omitting DSM testing and identifying which patients should continue to routinely undergo DSM testing.   Personalized shared decision-making regarding the use and role of DSM testing for ICD placement and replacement in congenital heart disease patients should continue.