Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jun 5.
Published in final edited form as: Circulation. 2018 Feb 20;137(23):2463–2477. doi: 10.1161/CIRCULATIONAHA.117.032167

Ventricular Fibrillation Conversion Testing After Implantation of a Subcutaneous Implantable Cardioverter Defibrillator: A Report from the National Cardiovascular Data Registry

Daniel J Friedman 1,2, Craig S Parzynski 3, E Kevin Heist 4, Andrea M Russo 5, Joseph G Akar 3, James V Freeman 3, Jeptha P Curtis 3, Sana M Al-Khatib 1,2
PMCID: PMC5988932  NIHMSID: NIHMS945774  PMID: 29463509

Abstract

Background

Compared to transvenous (TV) implantable cardioverter defibrillators (ICD), subcutaneous (S) ICDs require a higher energy for effective defibrillation. Although ventricular fibrillation (VF) conversion testing (CT) is recommended after S-ICD implantation to ensure an adequate margin between the defibrillation threshold and maximum device output (80J), prior work found that adherence to this recommendation is declining.

Methods

We studied first time S-ICD recipients (between September 28, 2012 and April 1, 2016) in the National Cardiovascular Database Registry ICD Registry to determine: predictors of use of CT, predictors of an insufficient safety margin (ISM, defined as VF conversion energy >65J) during testing, and in-hospital outcomes associated with use of CT. Multivariable logistic regression analysis was used to predict use of CT and ISM. Inverse probability weighted logistic regression analysis was used to examine the association between use of CT and in-hospital adverse events including death.

Results

CT testing was performed in 70.7% (n=5,624) of 7,960 S-ICD patients. Although deferral of CT was associated with several patient characteristics (including increased body mass index, increased body surface area, severely reduced ejection fraction, dialysis dependence, warfarin use, anemia, hypertrophic cardiomyopathy), the facility effect was comparatively more important (area under the curve for patient level versus generalized linear mixed model: 0.619 vs. 0.877). An ISM occurred in 6.9% (n=336) of 4,864 patients without a prior ICD and was more common among white patients and those with ventricular pacing on the pre-implant ECG, higher pre-implant blood pressure, larger body surface area, higher body mass index, and lower ejection fraction. A risk score was able to identify patients at low (<5%), medium (5-10%), and higher risk (>10%) for ISM. CT testing was not associated with a composite of in-hospital complications including death.

Conclusions

Use of CT testing after S-ICD implantation was driven by facility preference to a greater extent than patient factors and was not associated with a composite of in-hospital complications or death. ISM was relatively uncommon and is associated with several widely available patient characteristics. These data may inform ICD system selection and a targeted approach to CT.

Keywords: subcutaneous implantable cardioverter defibrillator, defibrillation, implantable cardioverter defibrillator

INTRODUCTION

The subcutaneous (S) implantable cardioverter defibrillator (ICD)(Boston Scientific, Natick, MA) is an entirely subcutaneous system that does not require intraprocedural vascular access or endovascular defibrillator leads or coils.1 The S-ICD has a novel mechanism of defibrillation and is associated with an increased energy requirement for defibrillation when compared to traditional transvenous ICDs.1 These factors, in association with an absence of data on the safety of forgoing VF conversion testing at the time of implant, have made VF conversion testing at the time of S-ICD implant a Class I (level of evidence C) recommendation.2 This is in stark contrast to the recommended approach for defibrillation threshold testing (DFT) testing at the time of transvenous ICD implant, where testing is not required routinely and is typically reserved for those at higher risk for inadequate safety margins (ISM).2

Prior research from the National Cardiovascular Data Registry (NCDR) ICD Registry has demonstrated a compelling need for additional research on VF conversion testing at the time of S-ICD implantation based on: very high rates of successful VF conversion testing at the time of S-ICD implant (92.7% and 99.8% at ≤65J and ≤80J, respectively); declining adherence to the Class I recommendation for use of VF conversion testing at the time of S-ICD implant; and increased incidence of peri-procedural cardiac arrest among S-ICD recipients (compared to transvenous ICD recipients) which could be related to VF conversion testing.3 Based on these knowledge gaps regarding VF conversion testing after S-ICD implantation, we performed an analysis of the NCDR ICD registry to understand factors associated with use of VF conversion testing, predictors of ISM at the time of VF conversion testing, and in-hospital outcomes associated with VF conversion testing.

METHODS

The data, analytic methods, and study materials cannot be made available to other researchers by the authors for purposes of reproducing the results or replicating the procedure, because the data are owned by the NCDR. The Yale University Human Investigation Committee approved the present analysis with waiver of informed consent.

Patient Characteristics

Baseline characteristics were obtained from the ICD Registry V2.1 Data Collection Form and included demographics, history and risk factors, diagnostic studies, and relevant pre-procedure hospitalization data. We considered all patients who underwent S-ICD implantation between September 28, 2012, the date of the Food and Drug Administration approval, and April 1, 2016, the date at which the NCDR transitioned to an updated case report form that did not include information on VF conversion testing. For the 1st aim (predicting use of VF conversion testing), we excluded patients who underwent S-ICD generator replacement, reportedly underwent upper limit of vulnerability testing (which is not feasible with the S-ICD), and those implanted outside of the United States. For the 2nd aim (predictors of ISM), we included the subset of the Aim 1 patients who underwent VF conversion testing, and additionally excluded those with a reported VF conversion testing result outside of the plausible range (<10J or >80J) and those with a prior ICD (or missing information regarding history of a prior ICD). For the 3rd aim (in-hospital outcomes), we included the subset of patients from the 1st aim who were admitted for the S-ICD procedure and excluded those with a prior ICD (or missing information regarding history of a prior ICD).

Predictors of interest

We considered several variables for the multivariable models, including demographics (age, sex, ethnicity, and race), anthropometric variables (height, weight, body mass index, and body surface area), hospital characteristics, and past medical history. GFR was calculated using the Modification of Diet in Renal Disease formula.4 Implausible clinical values for continuous variables were set to missing prior to the start of the analysis and later imputed. Both missing continuous variables and categorical variables were imputed using fully conditional specification prior to modeling.5

Outcomes

All outcomes were ascertained from the NCDR ICD Registry case report form. For the 1st aim, which sought to identify predictors of use of VF conversion testing after S-ICD implantation, the outcome of interest was use of VF conversion testing as reported in the case report form. For the 2nd aim, ISM was defined as when the lowest energy that resulted in conversion of ventricular fibrillation was >65J during the procedure. This definition was based on the widely held belief that a successful conversion at 65J represents an adequate safety margin relative to 80J, which is the only programmable output used outside of testing in the electrophysiology laboratory. Since device revision due to an ISM may have been performed prior to the lowest achieved VF conversion energy, the VF conversion value for this study best reflects the VF conversion energy at the final device position. The ICD Registry does not contain data regarding if multiple tests were performed and if device revisions were required to achieve the lowest reported value; as such, data on the frequency and outcomes associated with multiple testing and device revisions cannot be presented. For the 3rd aim, the primary endpoint was a composite of in-hospital adverse events consisting of: death, cardiac arrest, cardiac perforation, valve injury, hematoma requiring re-operation or blood transfusion, hemothorax, infection, lead dislodgement, myocardial infarction, pericardial tamponade, set screw problem, pneumothorax, transient ischemic attack or stroke, or urgent cardiac surgery.

Statistical Analysis

Aim1: Predictors of use of VF conversion Testing

We described the characteristics of patients, centers, and physicians associated with use versus non-use of VF conversion testing. Between group differences were tested using the Chi-square test for categorical variables and the Wilcoxon rank sum tests or t-tests for continuous variables. We employed multivariable logistic regression using Generalized Estimating Equations (GEE) accounting for clustering within facilities. Purposeful selection (as described by Hosmer and Lemeshow6) was used for patient level model selection. All variables in Table 1 were considered for inclusion except for year of implantation and giant cell myocarditis (n=1, precluding ability to generate an estimate). When a non-linear relationship between a continuous variable and use/non-use of VF conversion testing was observed, we plotted the relationship and binned the variable based on the observed relationship using clinically relevant thresholds. Generalized linear mixed models and hospital specific median odds ratios7 were later used to ascertain facility specific effects.

Table 1.

Patient, hospital, and physician characteristics for S-ICD patients who did and did not undergo DFT testing.

Total
(n=7,960)		 VF conversion Testing
(n=5,624)		 No VF conversion Testing
(n=2,336)
  Mean or absolute number Proportion or standard deviation Mean or absolute number Proportion or standard deviation Mean or absolute number Proportion or standard deviation
Age - Mean(SD) 53.0 15.4 52.6 15.3 54.1 15.5
Female 2,417 30.4 1,682 29.9 735 31.5
Race
 White non-Hispanic 4,813 60.5 3,524 62.7 1,289 55.2
 Black non-Hispanic 2,217 27.9 1,470 26.1 747 32.0
 Hispanic 573 7.2 380 6.8 193 8.3
 Other 357 4.5 250 4.5 107 4.6
Insurance Payor (not mutually exclusive)
 Private 4,748 59.7 3,398 60.4 1,350 57.8
 Medicare 3,468 43.6 2,360 42.0 1,108 47.4
 Medicaid 1,778 22.3 1,221 21.7 557 23.8
 Other 334 4.2 250 4.5 84 3.6
 None 236 3.0 172 3.1 64 2.7
Height, m - Mean(SD) 1.7 0.1 1.7 0.1 1.7 0.1
Weight, kg - Mean(SD) 88.3 23.4 88.3 22.8 88.4 24.7
Body Surface Area, m2 - Mean (SD) 29.5 6.9 29.5 6.8 29.7 7.2
Body Mass Index, kg/m2-Mean (SD) 2.0 0.3 2.0 0.3 2.0 0.3
QRS Duration, ms - Mean(SD) 103.6 20.3 103.5 20.1 103.9 20.8
PR Interval, ms - Mean(SD) 170.8 31.9 170.3 31.7 172.1 32.3
Non-Ischemic Dilated Cardiomyopathy 3,352 42.2 2,319 41.3 1,033 44.4
Ischemic HD 3,423 43.0 2,432 43.3 991 42.5
Atrial Fibrillation/Flutter 1,447 18.2 910 16.2 537 23.0
Ventricular Tachycardia 2,270 28.6 1,611 28.7 659 28.3
Preoperative Warfarin 1,310 16.5 826 14.7 484 20.7
 If yes-Held 947 72.5 588 71.3 359 74.5
 If yes-INR 1.6 0.6 1.6 0.6 1.6 0.7
NYHA Class
 I 2,029 25.7 1,493 26.7 536 23.1
 II 3,326 42.1 2,392 42.8 934 40.3
 III 2,425 30.7 1,628 29.1 797 34.4
 IV 129 1.6 77 1.4 52 2.2
Previous ICD 1,004 12.6 711 12.6 293 12.5
Previous Pacemaker 138 1.7 90 1.6 48 2.1
Previous PCI 2,162 27.2 1,573 28.0 589 25.3
Previous CABG 1,307 16.5 937 16.7 370 15.9
Diabetes 2,917 36.7 2,015 35.9 902 38.7
Previous MI 3,079 38.7 2,191 39.0 888 38.1
Structural Abnormalities
 Amyloidosis 5 0.1 2 0.0 3 0.1
 Ebsteins anomaly 5 0.1 3 0.1 2 0.1
 Transposition of the Great Vessels 23 0.3 12 0.2 11 0.5
 Giant Cell Myocarditis 1 0.0 1 0.0 0 0.0
 LV non-compaction 57 0.7 46 0.8 11 0.5
 Tetrology of Fallot 26 0.3 15 0.3 11 0.5
 Hypertrophic cardiomyopathy 428 5.4 339 6.1 89 3.8
 ARVD 44 0.6 36 0.6 8 0.3
 Common Ventricle 8 0.1 5 0.1 3 0.1
Syndromes Associated with SCD
 Long QT Syndrome 334 4.2 239 4.3 95 4.1
 Short QT Syndrome 6 0.1 3 0.1 3 0.1
 Brugada Syndrome 129 1.6 92 1.6 37 1.6
 Catecholaminergic Polymorphic VT 14 0.2 9 0.2 5 0.2
 Idiopathic VF 117 1.5 87 1.6 30 1.3
Chronic Lung Disease 1,291 16.2 882 15.7 409 17.5
Sleep Apnea
 No 4,612 58.1 3,248 57.9 1,364 58.6
 Yes 1,094 13.8 767 13.7 327 14.1
 Not Assessed 2,237 28.2 1,600 28.5 637 27.4
Cerebrovascular Disease 937 11.8 637 11.3 300 12.9
Cardiac Arrest 1,582 19.9 1,149 20.5 433 18.6
Hypertension 5,613 70.6 3,919 69.8 1,694 72.6
Syncope 1,310 16.5 940 16.7 370 15.9
CHF Duration >9 months (among those with HF) 3,991 68.5 2,762 68.1 1,229 69.3
AV Conduction
 2nd degree block 16 0.2 8 0.1 8 0.3
 3rd degree block 13 0.2 7 0.1 6 0.3
 Ventricular Pacing 67 0.8 46 0.8 21 0.9
EF % - Mean(SD) 31.8 14.3 32.5 14.4 30.1 14.0
GFR, ml/min/1.73 m2
 ≥60 4,807 61.1 3,541 63.7 1,266 55.0
 30-59 1,258 16.0 869 15.6 389 16.9
 15-29 254 3.2 171 3.1 83 3.6
 <15 including those on dialysis 1,543 19.6 978 17.6 565 24.5
Dialysis 1,430 18.0 917 16.3 513 22.0
BUN mg/dL - Median (25th-75th) 19.0 (14.00-29.00) 18.0 (13.00-27.00) 20.0 (14.00-32.00)
Hemoglobin, g/dL - Mean(SD) 12.7 2.2 12.8 2.1 12.4 2.2
Potassium, mEq/L- Mean(SD) 4.2 0.5 4.2 0.5 4.3 0.5
Sodium, mEq/L - Mean(SD) 138.5 3.2 138.5 3.2 138.4 3.4
Systolic BP, mmHg-Mean(SD) 127.2 22.6 127.2 22.6 127.1 22.7
Diastolic BP, mmHg-Mean(SD) 73.4 14.1 73.6 14.1 72.9 14.0
Reason for Admission
 Procedure 6,043 76.0 4,349 77.4 1,694 72.6
 Heart Failure 420 5.3 248 4.4 172 7.4
 Other Cardiac 1,254 15.8 874 15.6 380 16.3
 Non-Cardiac 238 3.0 149 2.7 89 3.8
EP Operator Training
 Board Certified 6,230 78.3 4,475 79.6 1,755 75.1
 Surgeon 52 0.7 32 0.6 20 0.9
 Other 1,389 17.5 917 16.3 472 20.2
 None 244 3.1 165 2.9 79 3.4
 Unknown 45 0.6 35 0.6 10 0.4
Census Region
 Midwest Region 1,995 25.1 1,465 26.1 530 22.7
 Northeast Region 1,871 23.5 1,267 22.5 604 25.9
 South Region 2,916 36.6 2,042 36.3 874 37.4
 West Region 1,178 14.8 850 15.1 328 14.0
Is a teaching hospital 5,271 66.2 3,652 64.9 1,619 69.3
Is a public hospital 4,150 52.1 3,051 54.3 1,099 47.1
# of Patient Beds - Median (25th-75th) 533.0 (369.00-714.00) 515.0 (372.00-710.50) 554.0 (366.00-751.00)
# of Patient Beds
 ≤100 132 1.7 77 1.4 55 2.4
 101-500 3,561 44.7 2,652 47.2 909 38.9
 501+ 4,267 53.6 2,895 51.5 1,372 58.7
Community
 Rural 515 6.5 323 5.7 192 8.2
 Suburban 2,111 26.5 1,468 26.1 643 27.5
 Urban 5,334 67.0 3,833 68.2 1,501 64.3
Profit Type
 Government 74 0.9 56 1.0 18 0.8
 Private/Community 5,320 66.8 3,796 67.5 1,524 65.2
 University 2,566 32.2 1,772 31.5 794 34.0
Implant Year
 2012 73 0.9 64 1.1 9 0.4
 2013 339 4.3 272 4.8 67 2.9
 2014 2,501 31.4 1,891 33.6 610 26.1
 2015 3,913 49.2 2,643 47.0 1,270 54.4
 2016 1,134 14.3 754 13.4 380 16.3
Open in a new tab

BMI=body mass index, CABG=coronary artery bypass grafting, BUN=blood urea nitrogen, CHF=congestive heart failure, EF=ejection fraction, EP=Electrophysiology, GFR= glomerular filtration rate ICD=implantable cardioverter defibrillator, INR=international normalized ratio, NYHA=New York Heart Association, MI=myocardial infarction, PCI=percutaneous coronary intervention, SD=standard deviation, VF=ventricular fibrillation, VT=ventricular tachycardia,

Aim 2: Predictors of ISM

We described the characteristics of patients, centers, and physicians associated with sufficient vs. ISM after S-ICD implantation. Between group differences and non-linear continuous variables were handled as per Aim 1. We created patient level multivariable logistic regression models (using GEE to account for clustering within facility) to identify independent predictors of an ISM at VF conversion testing. We considered all variables from Aim 1 after additionally excluding several variables with low prevalence: amyloidosis, Ebsteins anomaly, transposition of the great vessels, common ventricle, and short QT syndrome. Beta coefficients derived from the model were subsequently used to generate a weighted risk score.

We included sensitivity analyses assessing the association between use of amiodarone and sotalol at time of discharge and ISM, given the known impact of these drugs on the defibrillation threshold of transvenous ICDs. These variables were not considered in the main model because the case report form only collects discharge medication and it is possible that these medications could have been started or stopped after VF conversion testing.

Aim 3: The association between VF conversion testing and in-hospital events

In order to determine the risk adjusted association between use of VF conversion testing and in-hospital outcomes, we utilized generalized boosted models (GBM) to calculate inverse probability of treatment weights.8, 9 For each model, GBM fits a piecewise constant model to predict the probability of VF conversion testing. The model consists of many simple regression trees iteratively combined to create an overall piecewise constant function. The GBM models were fitted iteratively until the imbalance between groups was minimized. For our models, we set the maximum number of trees to be 20,000 and restricted interactions to two levels. Balance between groups was assessed using weighted standardized differences with values < 10% being considered sufficiently balanced.10, 11

We calculated weighted average treatment effects for each outcome defined above using weighted logistic regression with the outcome of interest as the dependent variable and VF conversion testing as the independent variable. Comparisons were summarized using odds ratios (ORs) with 95% confidence limits.

Analyses were performed using SAS (Version 9.4, SAS institute, Cary, NC) and R (Version 3.4.0, R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Predictors of Use of VF conversion Testing

A total of 8,190 patients underwent S-ICD implantation between September 28, 2012, and April 1, 2016. After excluding 60 patients who underwent S-ICD generator replacement, 169 patients who reportedly underwent upper limit of vulnerability testing (which is not possible with the S-ICD), and 1 patient who was implanted outside the United States, a total of 7,960 patients remained for analysis (Figure 1). Of these patients, 5,624 (70.7%) underwent VF conversion testing. The baseline characteristics of patients who did and did not undergo VF conversion testing are depicted in Table 1.

Figure 1.

Figure 1

Open in a new tab

A consort diagram depicting cohort derivation for the 3 study aims. Patients with a history of either S-ICD or transvenous ICD were excluded from the 2nd and 3rd aims.

In logistic regression analyses, deferral of VF conversion testing was independently associated with several patient characteristics (Table 2), including increased body mass index, increased body surface area, severely reduced ejection fraction (<20%), dialysis dependence, warfarin use, and anemia. VF conversion testing was more common in patients with prior coronary artery bypass grafting, hypertrophic cardiomyopathy, and implant at a public hospital.

Table 2.

Multivariable model predicting non-use of VF conversion *

  OR (95% CI)† p-value
Body Mass Index, kg/m2 (per 10 unit increase) 1.14(1.02-1.27) 0.0188
Body Surface Area, m2    
 <2.0 1.12(0.97-1.30) 0.1235
 ≥2.0 to <2.5 Reference  
 ≥2.5 + 1.60(1.17-2.19) 0.0033
PR Interval Not Obtainable 2.90(2.31-3.65) 0.0000
Preoperative Warfarin    
  No Reference  
  Yes-not held 1.29(0.97-1.73) 0.0838
  Yes-held 1.61(1.33-1.96) 0.0000
NYHA Class    
  I Reference  
  II 0.83(0.67-1.02) 0.0694
  III 1.02(0.82-1.27) 0.8357
  IV 1.44(0.89-2.34) 0.1400
Previous CABG 0.78(0.65-0.93) 0.0058
Transposition of the Great Vessels 3.62(1.32-9.91) 0.0124
Hypertrophic cardiomyopathy 0.67(0.48-0.94) 0.0187
Cardiac Arrest 0.76(0.63-0.91) 0.0037
EF, %    
  <20 1.76(1.24-2.50) 0.0016
  20-29 0.99(0.72-1.36) 0.9413
  30-39 0.84(0.61-1.15) 0.2730
  40-49 Reference  
  ≥50 1.10(0.79-1.51) 0.5820
GFR, ml/min/1.73 m2    
  ≥60 Reference  
  30-59 1.09(0.91-1.30) 0.3621
  15-29 0.97(0.68-1.39) 0.8867
  <15 including those on dialysis 1.38(1.14-1.66) 0.0008
Hemoglobin, g/dL    
  <9 2.03(1.49-2.75) 0.0000
  9-12.9 1.28(1.10-1.48) 0.0010
  13-16.9 1.00  
  ≥17 1.37(0.80-2.34) 0.2537
Potassium, mEq/L    
  <3.5 1.33(0.97-1.82) 0.0804
  3.5-4.4 1.00  
  ≥4.5 1.19(1.03-1.37) 0.0153
Public (vs. private) hospital 0.65(0.48-0.88) 0.0054
Open in a new tab
*

The area under the curve for the patient level model is 0.619, compared to 0.877 for the generalized linear mixed model. The hospital specific median odds ratio represents the median value of the odds ratio generated by comparing the rates of VF conversion testing in two randomly selected hospitals.7 The hospital specific median odds ratio for use of VF conversion was 5.16, emphasizing that facility preference is a very strong factor in determining use vs. non-use of VF conversion testing.

†

Odds ratio should be interpreted as the odds of not undergoing VF conversion testing

‡

CABG=coronary artery bypass grafting, EF=ejection fraction, GFR=glomerular filtration rate, NYHA=New York Heart Association

The area under the curve for the generalized linear mixed model that incorporated a facility specific effect was substantially greater than that for the patient level model only accounting for facility clustering (0.877 vs. 0.619) and the hospital specific median odds ratio was 5.16, demonstrating facility preference is a major determinant of VF conversion testing.

Predictors of ISM at DFT Testing

For this analysis, we additionally excluded patients with an improbable VF conversion testing value (n=56) or a prior ICD (or missing data regarding prior ICD)(n=704). Of the resulting 4,864 patients, a total of 336 (6.9%) were found to have an ISM; all patients had successful VF conversion at ≤80J. The baseline characteristics of patients who did and did not have ISM at time of VF conversion testing are depicted in Table 3. Two of the patients with an ISM required a separate procedure (during the index hospitalization) to revise the S-ICD.

Table 3.

Patient and physician characteristics among all S-ICD patients stratified by VF conversion testing result (sufficient vs. insufficient safety margin).

Total
(n=4,864)		 Sufficient
(n=4,528)		 Insufficient
(n=336)
Mean or absolute number Proportion or standard deviation Mean or absolute number Proportion or standard deviation Mean or absolute number Proportion or standard deviation
Age - Mean (SD) 52.36 15.38 52.44 15.40 51.30 15.03
Female 1,476 30.35 1,391 30.72 85 25.30
Race
 White non-Hispanic 3,009 61.86 2,787 61.55 222 66.07
 Black non-Hispanic 1,295 26.62 1,217 26.88 78 23.21
 Hispanic 339 6.97 318 7.02 21 6.25
 Other 221 4.54 206 4.55 15 4.46
Insurance Payor (not mutually exclusive)
 Private 2,944 60.53 2,746 60.64 198 58.93
 Medicare 1,965 40.40 1,826 40.33 139 41.37
 Medicaid 1,047 21.53 971 21.44 76 22.62
 Other 225 4.63 211 4.66 14 4.17
 None 159 3.27 148 3.27 11 3.27
Height, m - Mean (SD) 1.73 0.11 1.73 0.11 1.75 0.12
Weight, kg - Mean (SD) 88.13 22.63 87.21 22.05 100.63 26.45
Body Surface Area, m2 - Mean (SD) 2.01 0.27 2.00 0.26 2.14 0.29
Body Mass Index, kg/m2 - Mean (SD) 29.43 6.81 29.17 6.64 32.94 8.07
QRS Duration, ms - Mean (SD) 102.86 19.43 102.68 19.53 105.30 17.93
PR Interval, ms - Mean (SD) 169.58 31.21 169.31 31.19 173.42 31.29
Non-Ischemic Dilated Cardiomyopathy 2,064 42.50 1,893 41.87 171 50.89
Ischemic HD 2,069 42.54 1,930 42.62 139 41.37
Atrial Fibrillation / Flutter 721 14.83 659 14.56 62 18.45
Ventricular Tachycardia 1,227 25.24 1,147 25.35 80 23.81
Preoperative Warfarin 645 13.27 592 13.08 53 15.77
 If yes - held 448 69.57 413 69.88 35 66.04
 If yes - INR 1.63 0.57 1.64 0.58 1.55 0.48
NYHA Class
 I 1,255 25.94 1,190 26.43 65 19.40
 II 2,091 43.22 1,939 43.06 152 45.37
 III 1,436 29.68 1,320 29.31 116 34.63
 IV 56 1.16 54 1.20 2 0.60
Previous Pacemaker 52 1.07 45 0.99 7 2.08
Previous PCI 1,339 27.54 1,249 27.59 90 26.87
Previous CABG 764 15.73 730 16.14 34 10.12
Diabetes 1,734 35.66 1,609 35.55 125 37.20
Previous MI 1,862 38.30 1,736 38.36 126 37.50
Structural Abnormalities
 Amyloidosis 2 0.04 2 0.04 0 0.00
 Ebsteins anomaly 3 0.06 3 0.07 0 0.00
 Transposition of the Great Vessels 10 0.21 10 0.22 0 0.00
 Giant Cell Myocarditis 1 0.02 1 0.02 0 0.00
 LV non-compaction 43 0.89 40 0.89 3 0.90
 Tetrology of Fallot 13 0.27 12 0.27 1 0.30
 Hypertrophic cardiomyopathy 288 5.94 273 6.05 15 4.48
 ARVD 31 0.64 30 0.66 1 0.30
 Common Ventricle 5 0.10 5 0.11 0 0.00
Syndromes Associated with SCD
 Long QT Syndrome 204 4.20 189 4.18 15 4.46
 Short QT Syndrome 3 0.06 3 0.07 0 0.00
 Brugada Syndrome 81 1.67 79 1.75 2 0.60
 Catecholaminergic Polymorphic VT 8 0.16 8 0.18 0 0.00
 Idiopathic VF 72 1.48 65 1.44 7 2.08
Chronic Lung Disease 740 15.23 681 15.05 59 17.61
Sleep Apnea
 No 2,854 58.74 2,668 58.99 186 55.36
 Yes 637 13.11 574 12.69 63 18.75
 Not Assessed 1,368 28.15 1,281 28.32 87 25.89
Cerebrovascular Disease 538 11.07 502 11.09 36 10.75
Cardiac Arrest 953 19.60 890 19.66 63 18.75
Hypertension 3,380 69.52 3,135 69.25 245 73.13
Syncope 792 16.28 744 16.43 48 14.29
CHF Duration >9 months (among those with HF) 2,282 64.50 2,123 64.88 159 59.77
AV Conduction
 2nd degree block 5 0.10 5 0.11 0 0.00
 3rd degree block 7 0.14 7 0.15 0 0.00
 Ventricular Pacing 28 0.58 21 0.46 7 2.08
EF % - Mean (SD) 32.10 14.23 32.25 14.28 30.20 13.35
GFR, ml/min/1.73 m2
 ≥60 3,082 64.09 2,867 64.02 215 64.95
 30-59 723 15.03 670 14.96 53 16.01
 15-29 156 3.24 147 3.28 9 2.72
 <15 including those on dialysis 848 17.63 794 17.73 54 16.31
Dialysis 794 16.34 744 16.45 50 14.93
BUN – mg/dL Median (25th-75th) 18.00 (13.00-28.00) 18.00 (13.00-28.00) 18.00 (14.00-28.00)
Hemoglobin, g/dL - Mean(SD) 12.81 2.12 12.81 2.10 12.83 2.31
Potassium, mEq/L - Mean(SD) 4.24 0.48 4.23 0.47 4.25 0.50
Sodium, mEq/L - Mean(SD) 138.53 3.16 138.53 3.17 138.52 3.03
Systolic BP, mmHg - Mean(SD) 127.57 22.75 127.30 22.65 131.32 23.88
Diastolic BP, mmHg - Mean(SD) 73.80 14.09 73.68 14.06 75.36 14.35
Reason for Admission
 Procedure 3,725 76.65 3,464 76.57 261 77.68
 Heart Failure 226 4.65 212 4.69 14 4.17
 Other Cardiac 790 16.26 738 16.31 52 15.48
 Non-Cardiac 119 2.45 110 2.43 9 2.68
EP Operator Training
 Board Certified 3,856 79.28 3,600 79.51 256 76.19
 Surgeon 23 0.47 23 0.51 0 0.00
 Other 814 16.74 748 16.52 66 19.64
 None 137 2.82 128 2.83 9 2.68
 Unknown 34 0.70 29 0.64 5 1.49
Implant Year
 2012 50 1.03 45 0.99 5 1.49
 2013 209 4.30 202 4.46 7 2.08
 2014 1,630 33.51 1,520 33.57 110 32.74
 2015 2,317 47.64 2,150 47.48 167 49.70
 2016 658 13.53 611 13.49 47 13.99
Open in a new tab

BMI=body mass index, CABG=coronary artery bypass grafting, BUN=blood urea nitrogen, CHF=congestive heart failure, EP=Electrophysiology, ICD=implantable cardioverter defibrillator, INR=international normalized ratio, NYHA=New York Heart Association, MI=myocardial infarction, PCI=percutaneous coronary intervention, SD=standard deviation, VF=ventricular fibrillation, VT=ventricular tachycardia

ISM was more common among white patients and those with ventricular pacing on the pre-implant ECG, higher pre-implant blood pressure, larger body surface area, higher body mass index, and lower ejection fraction (Table 4). Prior coronary artery bypass grafting was associated with a lower likelihood of ISM. There was no association between ISM and age, sex, NYHA class, history of cardiac arrest, hypertrophic cardiomyopathy, or dialysis status. The area under the curve for the adjusted model was 0.679. A risk score was able to identify patients at low (<5% for scores 0-9), medium (5-10% for scores 10-16), and higher risk (>10% for scores ≥17) for ISM (Figure 2). Supplemental Table 1 details how the risk score was calculated for each patient.

Table 4.

Multivariable model predicting insufficient safety margin (AUC=0.679)

  Estimate SE OR (95% CI)* p-value overall p-value
Intercept −2.328 0.10      
Race       0.0150
 White non-Hispanic Ref 1.0    
 Black non-Hispanic −0.462 0.15 0.63(0.47-0.85) 0.0022  
 Hispanic −0.236 0.23 0.79(0.50-1.25) 0.3105  
 Other 0.041 0.28 1.04(0.60-1.81) 0.8829  
Body Surface Area, m2       0.0064
 <2.0 −0.402 0.13 0.67(0.52-0.86) 0.0019  
 2.0 to <2.5 Ref 1.0    
 ≥2.5 0.239 0.20 1.27(0.85-1.90) 0.2432  
Body Mass Index, kg/m2 (OR given for per 10 unit increase) 0.051 0.01 1.67(1.40- 1.99) 0.0000  
Previous CABG −0.609 0.18 0.54(0.38-0.78) 0.0009  
Ventricular Pacing on pre-implantation ECG 1.679 0.44 5.36(2.24-12.81) 0.0002  
EF % (OR given per 10 unit increase) −0.014 0.00 0.87(0.79-0.95) 0.0017  
Systolic BP (OR given per 10 unit increase) 0.007 0.00 1.08(1.02-1.13) 0.0036  
Open in a new tab

BP=blood pressure, CABG=coronary artery bypass grafting, EF=ejection fraction, OR = odds ratio

Figure 2.

Figure 2

Open in a new tab

A depiction of the frequency risk scores (Y axis on left) and observed rate of insufficient safety margin (Y axis on right) as a function of the risk score (X axis).

We performed sensitivity analyses to assess the association between discharge amiodarone (n=319) and sotalol (n=31) and an ISM. In a series of adjusted models, we observed no association between either amiodarone (OR 1.36, CI 0.91-2.04, p=0.13) or sotalol (OR 1.57, CI 0.50-4.97, p=0.44) and an ISM.

In-hospital outcomes and use vs. non-use of VF conversion testing

In an inverse probability weighted cohort, the rate of in-hospital complications was overall low, and there was no significant difference among those who did (0.84%) and did not (1.16%) undergo VF conversion testing (OR 0.72, CI 0.47-1.09, p = 0.11; reference group = no VF conversion testing; Table 5). Supplemental Table 2 depicts the study cohort before and after inverse probability weighting. There were no differences in in-hospital death by group. Cardiac arrest was less common among those who underwent VF conversion testing (OR 0.49, CI 0.28-0.87, p=0.02).

Table 5.

Adverse events by VF conversion testing status

  VF Conversion Testing No VF Converseion Testing
Outcome % % OR(95%CI) p-value
Any in-hospital complication 0.84 1.16 0.72(0.47-1.09) 0.1124
Death 0.16 0.14 1.16(0.40-3.33) 0.7828
Cardiac Arrest 0.37 0.76 0.49(0.28-0.87) 0.0152
Valve Injury 0.00 0.07 – –
Hematoma requiring re-operation or transfusion 0.26 0.09 3.03(0.96-10.00) 0.0581
Hemothorax 0.03 0.00 – –
Infection 0.05 0.00 – –
Lead Dislodgement 0.09 0.22 0.45(0.15-1.33) 0.1481
Myocardial Infarction 0.06 0.07 0.85(0.16-4.54) 0.8457
TIA or Stroke 0.02 0.07 0.37(0.04-3.13) 0.36
Urgent Cardiac Surgery 0.02 0.00 – –
Open in a new tab

No instances of cardiac perforation, pericardial tamponade, set screw problem, or pneumotheorax, or occurred in either group. = transient ischemic attack; VF = ventricular fibrillation

DISCUSSION

This study, which represents the largest reported cohort of S-ICD patients, has several notable findings. First, although deferral of VF conversion testing after S-ICD implant is modestly associated with several patient characteristics suggesting the patterns of performing this test are influenced by clinical factors, non-patient factors (e.g. hospital characteristics and preferences) are more strongly associated with performance of VF conversion testing. Second, several patient characteristics, including white race, increased BSA, increased BMI, and depressed EF, were associated with an ISM at the time of VF conversion testing. Age, sex, NYHA class, history of cardiac arrest, hypertrophic cardiomyopathy, and dialysis status were not associated with ISM. Third, VF conversion testing after S-ICD implantation appears safe, as it was not associated with a composite of in-hospital complications including death, but this analysis was limited by the small number of events.

DFT testing was performed routinely after transvenous ICD implant for years in order to ensure an adequate safety margin (usually 10J) between the DFT and the maximum device output. As the maximum device output increased and the defibrillation technology evolved over time, the perceived need to perform DFT testing declined, leading to a pivotal trial that confirmed that routine DFT testing was indeed not necessary for most patients with a primary prevention transvenous ICD.12 On average, the S-ICD has a DFT that is about 3x higher (and somewhat more variable) compared to a transvenous ICD (36.6±19.8J vs. 11.1±8.5J).1 The volume of subcutaneous fat between the defibrillator coil and chest wall,13 as well as position of the coil and pulse generator,1, 13 have been identified as important factors in achieving optimal DFTs. Accordingly, the maximum (and only) programmable output for the S-ICD is 80J, compared to ~35-40J for transvenous ICDs. Although early S-ICD studies suggested very high rates of successful conversion of VF1, 14, 15, there remained a relative paucity of data on rates of successful defibrillation and no data on whether or not it was safe to forgo VF conversion testing after S-ICD implantation, and so the expert consensus statement on ICD programming that was published following the approval of the S-ICD recommended VF conversion testing after S-ICD implantation as a Class I recommendation, based solely on expert consensus.2

Despite a Class I recommendation, the use of VF conversion testing after S-ICD implantation has been declining in the US, from 82.4% to 71.4% between 2012 and 2015.3 The current study sought to determine factors associated with VF conversion testing. We found that non-use of VF conversion testing was more common among patients who may have been at higher perceived risk for VF conversion testing complications (patients with increased BMI/BSA, severely reduced LVEF, dialysis dependence, and oral anticoagulation). The adjusted model that incorporated the site effects was more robust with an area under the curve of 0.877 (vs. 0.619 for the model with patient factors only), demonstrating that site factors appear proportionally more important than patient factors. Given the limits of registry data, we are unable to determine why facility preference is such a strong predictor of use of VF conversion testing, although it is plausible that cumulative experience regarding strategies that address ISM has decreased the perceived need for compulsory testing.

An ISM after S-ICD implantation appears to be uncommon. In our study, 6.9% of patients did not have successful VF conversion at ≤65J although all had successful VF conversion at ≤80J. These results are consistent with those reported from the Evaluation oF FactORs ImpacTing CLinical Outcome and Cost EffectiveneSS of the S-ICD (EFFORTLESS S-ICD) Registry; an early report16 from this registry reported 95% of patients had successful VF conversion at ≤65J and a more contemporary report with approximately twice as many patients reported this rate was 91.6%.17

Although there are several well established predictors of an ISM among transveous ICDs,18 limited data exist for the S-ICD. A recently published report from the S-ICD Post Approval Study (PAS) showed that among the 1,412 patients who underwent conversion testing, 95.6% of first shocks were successful at the final device position.19 Prior transvenous ICD extraction, increased height, and increased BMI were independently associated with first shock failure, while black race was associated with first shock success.19 These findings are concordant with some of our findings that white race and increased BMI were associated with an ISM at the time of VF conversion testing. In contrast to our study, PAS study did not identify a reduced EF (handled as a continuous variable) as a predictor of first shock failure. Based on recent computer modeling emphasizing the impact of subcutaneous fat on DFTs with the S-ICD,13 it is not surprising that increased BMI and BSA were both independent predictors of an ISM in the current analysis. Of note, BSA was not considered in the S-ICD PAS predictive model. In our study, height was collinear with, and less predictive than, BSA and therefore was not in the final model. Notably, BMI and BSA are not usually associated with ISM in transvenous systems.

There are several other notable differences in predictors of ISM among S-ICD vs. transvenous ICD patients. Dialysis dependent chronic kidney disease has been linked to ISM with transvenous ICDs but was not associated with an ISM in our study of S-ICDs; this is clinically relevant because the absence of endovascular leads has been hypothesized to be particularly beneficial among dialysis patients who are at a high risk for intravascular infections and dialysis access complications. Decreased age (<70 years), Hispanic ethnicity, and secondary prevention ICD indication have been associated with ISM among transvenous ICD patients18 but were not associated with ISM after S-ICD implantation. Thus, predictors of ISM vary by device type; these characteristics could be considered when selecting an ICD system or in applying a targeted approach to VF conversion testing.

After inverse probability weighting, VF conversion testing was not associated with a significant increase in the risk of any in-hospital complication reported to the NCDR. These results are consistent with the use of DFT testing in appropriately selected transvenous ICD patients.12 The rate of cardiac arrest was lower among patients who underwent VF conversion testing. Although it is plausible that VF conversion testing resulted in a higher likelihood of successful conversion of spontaneous ventricular tachycardia or fibrillation in the hospital after S-ICD implant, precluding the need for cardiopulmonary resuscitation and external defibrillation, it is more likely that intra-procedural cardiac arrests decreased the likelihood of performing VF conversion testing at the conclusion of the implant. However, these hypotheses are not possible to test as the ICD Registry case report form does not specify whether a peri-procedural cardiac arrest occurred during or after the ICD procedure. These results should be confirmed in an adequately powered prospective randomized trial.

Limitations

This study has several important limitations. This study was observational and retrospective in nature and use of VF conversion testing was not randomized. As with all registry studies, data may be prone to more inaccuracies and underreporting of complications, when compared to randomized trials. However, prior analyses have suggested >90% accuracy for data fields.20 The VF conversion value used for this analysis represents the lowest value during the procedure, which could have been achieved after an ISM led to an in-lab device revision, followed by repeat induction and testing. Thus, the data on rates and predictors of ISM best reflect the final device position. The ICD Registry case report form does not collect data on whether a device revision or reprogramming occurred during the initial procedure, and as such we are unable to report on the use of intraprocedural strategies to mitigate ISM. There were insufficient patients in this study to use separate derivation and validation cohorts when designing a model to predict ISM. The relatively small number of patients on amiodarone or sotalol may have reduced the power to detect an association between use of these drugs and an ISM. We did not have data on several potentially important predictors of an ISM, including electrode and pulse generator location,1, 13 chest wall dimensions, cardiac dimensions,21 and amount of subcutaneous fat below the electrode and pulse generator.13 Although the inverse probability cohort appeared well balanced based on assessment of standardized differences, we cannot rule out the possibility of residual confounding, which would most likely lead to an underestimate of risk associated with VF conversion testing.22 In-hospital complications were relatively infrequent, limiting the statistical power to detect a difference among patients who did and did not undergo VF conversion testing. We did not have longitudinal outcomes which are necessary to determine if VF conversion testing is able improve clinical outcomes by way of improving defibrillation of spontaneous ventricular arrhythmias. Finally, the analysis of in-hospital complications included only patients implanted during an elective hospitalization, potentially limiting generalizability to patients implanted during a hospitalization for another reason (e.g. heart failure exacerbation).

Clinical Implications

This study has several clinically relevant implications. Although facility preference appears to be the dominant factor regarding decision to perform VF conversion testing, several of the patient characteristics associated with non-use of VF conversion testing were also associated with risk for an ISM (e.g., BSA, BMI, and severely reduced LVEF). This suggests that the patients who may be most poised to benefit from VF conversion testing are less likely to undergo the procedure. This could potentially be related to greater perceived risk of performing DFT in these patients. The risk for ISM after S-ICD implant can be predicted with moderate accuracy using several readily available clinical variables. Understanding risk for ISM has the potential to not only lead to a targeted approach to VF conversion testing, but also improve ICD system selection. For example, our study results suggest that, when compared to a transvenous ICD, the S-ICD may be associated with a more favorable safety margin among several patient groups, including dialysis patients, blacks, Hispanics, younger patients, and men. In contrast, the S-ICD, when compared to the transvenous ICD, appears to be associated with a less favorable safety margin among patients with an increased BMI and increased BSA. Although the current study suggests that VF conversion testing may be safe in most patients after S-ICD implant, there are several patient groups that the results should not be generalized to, including patients with a very severely reduced LVEF, unrevascularizated coronary artery disease, or an intra-cardiac thrombus.

CONCLUSION

Although the use of VF conversion testing after S-ICD implantation is associated with several patient characteristics, facility preference appears to have a greater overall influence on use of testing. An ISM is relatively infrequent after S-ICD implantation and is associated with several clinical variables. VF conversion testing was not associated with increased in-hospital complications or death. The results from this study have important implications for ICD system selection (transvenous vs. subcutaneous) and the development of a personalized approach to VF conversion testing.

Supplementary Material

clean supplemental data file

CLINICAL PERSPECTIVE.

What is new?

  • -

    This study demonstrated that use versus non-use of VF conversion testing after S-ICD implantation in the US is more related to physician preference than patient characteristics

  • -

    This study identified several patient characteristics associated with an insufficient defibrillation safety margin among S-ICD recipients, including increased body mass index, severely decreased ejection fraction, white race, and ventricular pacing on the pre-implantation ECG.

  • -

    Use of VF conversion testing after S-ICD implant was not associated with a composite of in-hospital complications or death.

What are the clinical implications?

  • -

    An understanding of risk for insufficient defibrillation safety margin after S-ICD implant may be useful for ICD selection (transvenous vs. subcutaneous) and a targeted approach for the use of VF conversion testing.

  • -

    VF conversion testing after S-ICD implant does not appear to be associated with excess risk of adverse events in properly selected patients.

Acknowledgments

Partners and Sponsors: ICD Registry™ is an initiative of the American College of Cardiology with partnering support from the Heart Rhythm Society.

Disclaimer: This research was supported by the American College of Cardiology’s NCDR. The views expressed in this manuscript represent those of the author(s), and do not necessarily represent the official views of the NCDR or its associated professional societies identified at CVQuality.ACC.org/NCDR.

Funding Source: This study was funded by the National Cardiovascular Data Registry (NCDR). Dr. Friedman has received funding via the National Institutes of Health T 32 training grant HL069749.

Disclosures: Dr. Curtis owns stock in Medtronic (significant), receives research funding from Boston Scientific (significant), and receives salary support from the American College of Cardiology to provide data analytic services (significant). Dr. Freeman reports serving on an Advisory Board from Janssen Pharmaceuticals (modest). Dr. Friedman has received educational grants from Boston Scientific (modest) and St. Jude (modest), research grants from the National Cardiovascular Data Registry (significant), and is funded by the National Institutes of Health T 32 training grant HL069749 (significant). Dr. Heist is a consultant for Biotronik (modest), Boston Scientific (modest), Pfizer (significant), and Abbott (modest), and has received research grants from Boston Scientific (significant) and Abbott (significant). Dr Russo received research support from Boston Scientific and Medtronic and honoraria or consulting fees from Biotronik, Boston Scientific, Medtronic, and St Jude. Mr. Parzynski receives salary support from the American College of Cardiology to provide data analytic services (moderate).The remaining authors report no disclosures.

References

  • 1.Bardy GH, Smith WM, Hood MA, Crozier IG, Melton IC, Jordaens L, Theuns D, Park RE, Wright DJ, Connelly DT, Fynn SP, Murgatroyd FD, Sperzel J, Neuzner J, Spitzer SG, Ardashev AV, Oduro A, Boersma L, Maass AH, Van Gelder IC, Wilde AA, van Dessel PF, Knops RE, Barr CS, Lupo P, Cappato R, Grace AA. An entirely subcutaneous implantable cardioverter-defibrillator. The New England journal of medicine. 2010;363:36–44. doi: 10.1056/NEJMoa0909545. [DOI] [PubMed] [Google Scholar]
  • 2.Wilkoff BL, Fauchier L, Stiles MK, Morillo CA, Al-Khatib SM, Almendral J, Aguinaga L, Berger RD, Cuesta A, Daubert JP, Dubner S, Ellenbogen KA, Mark Estes NA, 3rd, Fenelon G, Garcia FC, Gasparini M, Haines DE, Healey JS, Hurtwitz JL, Keegan R, Kolb C, Kuck KH, Marinskis G, Martinelli M, McGuire M, Molina LG, Okumura K, Proclemer A, Russo AM, Singh JP, Swerdlow CD, Teo WS, Uribe W, Viskin S, Wang CC, Zhang S. 2015 HRS/EHRA/APHRS/SOLAECE expert consensus statement on optimal implantable cardioverter-defibrillator programming and testing. Heart rhythm : the official journal of the Heart Rhythm Society. 2016;13:e50–86. doi: 10.1016/j.hrthm.2015.11.018. [DOI] [PubMed] [Google Scholar]
  • 3.Friedman DJ, Parzynski CS, Varosy PD, Prutkin JM, Patton KK, Mithani A, Russo AM, Curtis JP, Al-Khatib SM. Trends and In-Hospital Outcomes Associated With Adoption of the Subcutaneous Implantable Cardioverter Defibrillator in the United States. JAMA Cardiol. 2016;1:900–911. doi: 10.1001/jamacardio.2016.2782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130:461–470. doi: 10.7326/0003-4819-130-6-199903160-00002. [DOI] [PubMed] [Google Scholar]
  • 5.van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res. 2007;16:219–242. doi: 10.1177/0962280206074463. [DOI] [PubMed] [Google Scholar]
  • 6.Hosmer DW, Lemeshow S, Sturdivant RX. Applied Logistic Regression. 3rd. John Wiley & Sons; 2013. [Google Scholar]
  • 7.Merlo J, Chaix B, Ohlsson H, Beckman A, Johnell K, Hjerpe P, Rastam L, Larsen K. A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena. J Epidemiol Community Health. 2006;60:290–297. doi: 10.1136/jech.2004.029454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Lee BK, Lessler J, Stuart EA. Improving propensity score weighting using machine learning. Stat Med. 2010;29:337–346. doi: 10.1002/sim.3782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.McCaffrey DF, Ridgeway G, Morral AR. Propensity score estimation with boosted regression for evaluating causal effects in observational studies. Psychol Methods. 2004;9:403–425. doi: 10.1037/1082-989X.9.4.403. [DOI] [PubMed] [Google Scholar]
  • 10.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28:3083–3107. doi: 10.1002/sim.3697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Normand ST, Landrum MB, Guadagnoli E, Ayanian JZ, Ryan TJ, Cleary PD, McNeil BJ. Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores. J Clin Epidemiol. 2001;54:387–398. doi: 10.1016/s0895-4356(00)00321-8. [DOI] [PubMed] [Google Scholar]
  • 12.Healey JS, Hohnloser SH, Glikson M, Neuzner J, Mabo P, Vinolas X, Kautzner J, O’Hara G, VanErven L, Gadler F, Pogue J, Appl U, Gilkerson J, Pochet T, Stein KM, Merkely B, Chrolavicius S, Meeks B, Foldesi C, Thibault B, Connolly SJ, Shockless IEi Cardioverter defibrillator implantation without induction of ventricular fibrillation: a single-blind, non-inferiority, randomised controlled trial (SIMPLE) Lancet. 2015;385:785–791. doi: 10.1016/S0140-6736(14)61903-6. [DOI] [PubMed] [Google Scholar]
  • 13.Heist EK, Belalcazar A, Stahl W, Brouwer TF, Knops RE. Determinants of Subcutaneous Implantable Cardioverter-Defibrillator Efficacy: A Computer Modeling Study. J Am Coll Cardiol EP. 2017;3:405–414. doi: 10.1016/j.jacep.2016.10.016. [DOI] [PubMed] [Google Scholar]
  • 14.Weiss R, Knight BP, Gold MR, Leon AR, Herre JM, Hood M, Rashtian M, Kremers M, Crozier I, Lee KL, Smith W, Burke MC. Safety and efficacy of a totally subcutaneous implantable-cardioverter defibrillator. Circulation. 2013;128:944–953. doi: 10.1161/CIRCULATIONAHA.113.003042. [DOI] [PubMed] [Google Scholar]
  • 15.Burke MC, Gold MR, Knight BP, Barr CS, Theuns DA, Boersma LV, Knops RE, Weiss R, Leon AR, Herre JM, Husby M, Stein KM, Lambiase PD. Safety and Efficacy of the Totally Subcutaneous Implantable Defibrillator: 2-Year Results From a Pooled Analysis of the IDE Study and EFFORTLESS Registry. Journal of the American College of Cardiology. 2015;65:1605–1615. doi: 10.1016/j.jacc.2015.02.047. [DOI] [PubMed] [Google Scholar]
  • 16.Lambiase PD, Barr C, Theuns DA, Knops R, Neuzil P, Johansen JB, Hood M, Pedersen S, Kaab S, Murgatroyd F, Reeve HL, Carter N, Boersma L, Investigators E. Worldwide experience with a totally subcutaneous implantable defibrillator: early results from the EFFORTLESS S-ICD Registry. European heart journal. 2014;35:1657–1665. doi: 10.1093/eurheartj/ehu112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Boersma L, Barr C, Knops R, Theuns D, Eckardt L, Neuzil P, Scholten M, Hood M, Kuschyk J, Jones P, Duffy E, Husby M, Stein K, Lambiase PD, Group EI Implant and Midterm Outcomes of the Subcutaneous Implantable Cardioverter-Defibrillator Registry: The EFFORTLESS Study. Journal of the American College of Cardiology. 2017;70:830–841. doi: 10.1016/j.jacc.2017.06.040. [DOI] [PubMed] [Google Scholar]
  • 18.Hsu JC, Marcus GM, Al-Khatib SM, Wang Y, Curtis JP, Sood N, Parker MW, Kluger J, Lampert R, Russo AM. Predictors of an inadequate defibrillation safety margin at ICD implantation: insights from the National Cardiovascular Data Registry. Journal of the American College of Cardiology. 2014;64:256–264. doi: 10.1016/j.jacc.2014.01.085. [DOI] [PubMed] [Google Scholar]
  • 19.Gold MR, Aasbo JD, El-Chami MF, Niebauer M, Herre J, Prutkin JM, Knight BP, Kutalek S, Hsu K, Weiss R, Bass E, Husby M, Stivland TM, Burke MC. Subcutaneous implantable cardioverter-defibrillator Post-Approval Study: Clinical characteristics and perioperative results. Heart rhythm : the official journal of the Heart Rhythm Society. 2017;14:1456–1463. doi: 10.1016/j.hrthm.2017.05.016. [DOI] [PubMed] [Google Scholar]
  • 20.Messenger JC, Ho KK, Young CH, Slattery LE, Draoui JC, Curtis JP, Dehmer GJ, Grover FL, Mirro MJ, Reynolds MR, Rokos IC, Spertus JA, Wang TY, Winston SA, Rumsfeld JS, Masoudi FA. The National Cardiovascular Data Registry (NCDR) Data Quality Brief: the NCDR Data Quality Program in 2012. Journal of the American College of Cardiology. 2012;60:1484–1488. doi: 10.1016/j.jacc.2012.07.020. [DOI] [PubMed] [Google Scholar]
  • 21.Bonnes JL, Westra SW, Bouwels LH, MJ DEB, Brouwer MA, Smeets JL. Risk Factors for Inadequate Defibrillation Safety Margins Vary With the Underlying Cardiac Disease: Implications for Selective Testing Strategies. J Cardiovasc Electrophysiol. 2016;27:587–593. doi: 10.1111/jce.12940. [DOI] [PubMed] [Google Scholar]
  • 22.Russo AM, Wang Y, Al-Khatib SM, Curtis JP, Lampert R. Patient, physician, and procedural factors influencing the use of defibrillation testing during initial implantable cardioverter defibrillator insertion: findings from the NCDR(R) Pacing Clin Electrophysiol. 2013;36:1522–1531. doi: 10.1111/pace.12248. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

clean supplemental data file
NIHMS945774-supplement-clean_supplemental_data_file.pdf (201.7KB, pdf)

RESOURCES