Abstract
Background
Implantable cardioverter-defibrillators (ICDs) reduce the risk of sudden cardiac death in patients with left ventricular dysfunction. While short-term mortality benefit of ICD insertion has been established in landmark randomized controlled trials, little is known about the long-term outcomes of patients with ICDs in clinical practice. In this paper, we describe the long-term survival of patients following de novo ICD implantation for primary prevention in clinical practice and determine the factors which help predict survival after ICD implant.
Methods
Retrospective population-based study of all patients receiving a de novo ICD for primary prevention in Ontario, Canada from 2007 to 2011 using the Ontario ICD Database housed within ICES. Simple random selection was used to split the population into a derivation and internal validation cohort in a ratio of 2:1. Cox proportional hazards regression was used to determine predictors of interest and predict 10-year survival, model performance was assessed using calibration and validation.
Results
In the derivation cohort (n = 3399), mean age was 65.3 years (standard deviation [SD] = 11.0), 664 patients were female (19.5 %) and 2344 patients (69.0 %) had ischemic cardiomyopathy. Ten year survival was 45.7 % (95 % confidence interval [CI] 44.0 %–47.4 %). The final prediction model included age, sex, disease factors (ischemic vs nonischemic cardiomyopathy, left ventricular ejection fraction) and patient factors (symptoms, comorbidities), and biomarkers at the time of ICD assessment. This model had good discrimination and calibration in derivation (0.79, 95 % CI 0.77, 0.81) and validation samples (0.78, 95 % CI 0.76, 0.79).
Conclusions
A combination of demographic and clinical factors determined at baseline can be used to predict 10-year survival in patients with implantable cardioverter-defibrillators with good accuracy. Our findings help to identify individuals at risk of long-term mortality and may be useful in targeting future prevention strategies to enhance longevity in this high-risk population.
Keywords: Implantable cardioverter defibrillator, Heart failure, Cardiomyopathy, Long-term follow-up, Long-term outcomes, Population-level research
1. Introduction
Implantable cardioverter defibrillators (ICDs) are recommended to reduce the risk of sudden cardiac death (SCD) for patients at risk [[1], [2], [3]]. This includes patients with prior symptomatic arrhythmias for secondary prevention, but also patients with left ventricular dysfunction at high risk of SCD without prior cardiac events for primary prevention. Initial landmark randomized controlled trials demonstrated a reduction in arrhythmic death and all-cause mortality in patients at risk with reduced left ventricular ejection fraction [[4], [5], [6], [7], [8], [9]]. Subsequent studies have found that the addition of cardiac resynchronization therapy (CRT) further improves outcomes in symptomatic heart failure patients with a low ejection fraction and QRS prolongation [10]. Current guidelines recommend the use of ICDs in patients with persistent LV dysfunction (LVEF ≤30 %) after 3 months of optimal medical therapy for NICM and 3 months after revascularization or 40 days after MI for ICM [[1], [2], [3]].
However, ICD recipients in clinical practice are often older with a greater burden of cardiac and non-cardiac comorbidities [11,12]. These patients are often at a higher risk of complications after ICD implant as well as mortality from competing noncardiac comorbidities. The care of ICD recipients continues to carry significant personal, clinical, and economic burden.
Previous observational studies have found death rates ranging from 12 to 19 % at 3 years and 25–36 % at 5 years after ICD implant [[13], [14], [15]]. However, population-level assessment of predictors of long-term outcomes remains limited and few studies report outcomes at 10-years post ICD insertion. Understanding the predictors of long-term outcome in patients after ICD implantation can inform shared decision making between patients and clinicians when deciding whether ICD therapy is appropriate. The aim of this study was to describe the 10-year survival and determine the predictors of survival for adult patients receiving ICD therapy for primary prevention in Ontario, Canada using population-based data.
2. Methods
2.1. Patient sample
This study included patients undergoing de novo ICD implantation for primary prevention who were registered in the Ontario ICD Database [16] between February 2007 and March 2011. In brief, the Ontario ICD database was a prospective registry mandated by the Ontario Ministry of Health and Long-Term Care of all patients (age≥18 years) undergoing evaluation for defibrillator implantation in Ontario, Canada. Patient data were collected at initial evaluation, time of defibrillator implant and device-clinic follow-up. The Ontario ICD database was linked to other databases housed within ICES (formerly Institute for Clinical Evaluative Sciences). These datasets were linked using unique encoded identifiers and analyzed at ICES. As a prescribed entity under the Ontario health information privacy legislation housed within ICES (section 45 of Ontario's Personal Health Information Privacy Act), patient data were collected without consent thus minimizing participation bias [17]. The use of these data was authorized under section 45 of Ontario's Personal Health Information Protection Act, which does not require review by a research ethics board. However, we obtained additional ethics approval from Queen's University.
2.2. Data sources
Data from the Ontario ICD database were collected by the local electrophysiologist and trained research coordinator and entered into a secure, firewall- and password-protected web-based registry at ICES. Data on patient characteristics, defibrillator indication, left ventricular ejection fraction, comorbidities, medications, and biomarkers were collected. Data quality was assessed using (1) regular review to ensure accuracy of coded data, (2) automated notification of uncoded data, (3) logic check of key data elements, and (4) random site audits of data reliability [16].
2.3. Outcomes
The primary outcome of all-cause mortality was identified using the Registered Persons Database at ICES. Vital status was available for all study patients. Patients were censored alive at 10-years follow-up and if Ontario Health Insurance Plan eligibility expired prior to the end of the follow-up period (n = 1445).
2.4. Predictor variables
Patient factors including age, sex, comorbidities, medications, and biomarkers were determined from the Ontario ICD database at the time of ICD implant assessment. Relevant comorbidities include prior PCI or CABG, prior MI, atrial fibrillation, diabetes, hypertension, peripheral vascular disease, stroke or TIA, dialysis-dependent chronic kidney disease (CKD), and chronic obstructive pulmonary disease (COPD). Relevant medications included angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), beta-blockers, furosemide, aspirin, clopidogrel, oral anticoagulation, digoxin, amiodarone, and other class II antiarrhythmic medications. Relevant biomarkers include QRS duration, estimated glomerular filtration rate (eGFR), hemoglobin, and sodium. Overall comorbidity burden was determined using The Johns Hopkins ACG® System Version 9.0. Patient frailty was estimated using the Hospital Frailty Risk Score [18], which used a weighted combination of 109 ICD-9 and/or ICD-10 diagnostic codes recorded in the preceding 2 years.
Device factors including device type, right ventricular (RV) lead fixation method and device manufacturer were determined using the Ontario ICD database. Provider factors including physician procedure volume in the preceding fiscal year, cardiologist vs. non-cardiologist, Canadian medical graduate status, and years since medical school graduation were determined using the ICES physician database (IPDB).
Surrogate measures of patient socioeconomic status were determined using the Ontario Marginalization Index and neighborhood income quintile, as previously described for other population-level studies [19]. Geographical distance from patient's resident to ICD implantation center was calculated using the automated geographic coding based on Statistics Canada's postal code conversion files using each person's residence postal code [19].
2.5. Statistical analysis
Baseline characteristics of the study population were examined using descriptive statistics. Derivation and internal validation cohorts were created using a random 2:1 split of the sample [20]. Potential predictors were included in a multivariable Cox regression model using a backwards variable elimination with retention of variables based on clinical and statistical significance with a generous cut-off of p-value <0.2. All multivariable models included age, sex, and disease etiology because of the potential of these variables to account for survival after ICD implant. Predictors were then categorized into groups (device factors, comorbidities, provider factors etc.) and added in a stepwise fashion into the prediction model.
Continuous variables were examined using cubic spline analysis to determine the strength and shape of association with death. Variables at risk of non-linearity were categorized using clinical significance and thresholds identified using cubic spline. Variables with a large amount of missing data (>10 %) were excluded from the regression model (physician implant volume, left ventricular end diastolic dimension LVEDD), but were included in a sensitivity analysis after single imputation of missing variables. Addition of device factors, provider factors, non-clinical surrogates of patient comorbidity and frailty burden, interactions between sex and clinically relevant predictor variables, and allowing maximum follow-up were also included in the sensitivity analyses.
Model fit was assessed for each model using the Akaike information criterion, calibration, and discrimination indices. Cox proportional hazard assumptions were met in the final model. Discrimination was determined with the construction of a receiver operator characteristic curve using bootstrapping of 100 replicate samples with replacement and calculation of the area under the curve (AUC). Calibration was visually examined with a plot of the observed proportions of death for 10 artificially conducted predicted risk groups of equal size using a previously described SAS macro [21].
3. Results
3.1. Patient characteristics
Among 7947 patients referred for de novo primary prevention ICD insertion, 2015 refused the device and 156 were replacement procedures. Patients were excluded if they were un-linkable to administrative databases or had specialized indications for ICDs including hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, Brugada syndrome, long QT syndrome or congenital heart disease (n = 679). The final cohort contained 5097 patients, which was split into a derivation (n = 3399) and validation cohort (n = 1698) using random selection at a ratio of 2:1. The patient flow diagram is shown in Fig. 1.
Fig. 1.
Patient flow diagram
Abbreviations: ARVC, arrhythmogenic right ventricular cardiomyopathy; HCM, hypertrophic cardiomyopathy, LQTS, long QT syndrome; ICD, implantable cardioverter defibrillator.
The final derivation cohort was comprised of 2735 men (80.5 %), mean age was 65.28 years (SD 11.0 years), and primary disease indication was ischemic cardiomyopathy in 2344 (69.0 %) patients. Devices implanted included 1442 (42.4 %) single chamber, 782 (23.0 %) dual chamber, and 1173 (34.5 %) cardiac resynchronization therapy defibrillator devices. Demographic and clinical characteristics of these patients are reported in Table 1.
Table 1.
Baseline characteristics of patients undergoing ICD implantation for primary prevention from the Ontario ICD Registry.
Abbreviations: ACEI = angiotensin converting enzyme inhibitor; ACG = Adjusted Clinical Group; ARB = angiotensin receptor blocker; BUN = blood urea nitrogen; CABG = coronary artery bypass graft; CCS = Canadian Cardiovascular Society; CRT-D = cardiac resynchronization therapy - defibrillator; eGFR = estimated glomerular filtration rate; HF = heart failure; LVEDD = left ventricular end diastolic diamension; LVESD = left ventricular end systolic dimension; LVEF = left ventricular ejection fraction; NYHA = New York Heart Association; ICD = implantable cardioverter defibrillator; PCI = percutaneous coronary intervention; RV = right ventricle; std = standard deviation.
| Characteristic | Derivation cohort n = 3399 | Validation Cohort n = 1698 | |
|---|---|---|---|
| Demographic | Age, mean (std) | 65.28 (11.00) | 65.55 (10.86) |
| Female Sex | 664 (19.54 %) | 346 (20.4 %) | |
| Summary Score Median, mean (std) | 3.10 (0.77) | 3.13 (0.77) | |
| Neighborhood income quintile 1-2 | 1336 (39.30 %) | 699 (41.17 %) | |
| Neighborhood income quintile 3-5 | 2053 (60.40 %) | 992 (58.42 %) | |
| Rural | 483 (14.21 %) | 235 (13.84 %) | |
| Distance from residence to ICD center in km, mean (std) | 76.49 (151.68) | 69.50 (121.58) | |
| Hospital Frailty Risk Score, mean (std) | 1.55 (2.93) | 1.54 (3.00) | |
| John Hopkins ACG, mean (std) | 8.90 (3.37) | 8.87 (3.44) | |
| Primary Disease Indication | Nonischemic cardiomyopathy | 1023 (30.10 %) | 504 (29.68 %) |
| Ischemic cardiomyopathy | 2344 (68.96 %) | 1178 (69.38 %) | |
| Other cardiac disease | 32 (0.94 %) | 16 (0.94 %) | |
| Cardiac History | LVEF≤20 % | 760 (23.11 %) | 341 (20.74 %) |
| LVEF (20–30 %] | 1878 (57.12 %) | 953 (57.97 %) | |
| LVEF (30–35 %] | 402 (12.23 %) | 216 (13.14 %) | |
| LVEF >35 % | 248 (7.54 %) | 134 (8.15 %) | |
| Prior myocardial infarction | 2002 (58.90 %) | 983 (57.89 %) | |
| Prior PCI or CABG | 1698 (49.96 %) | 818 (48.17 %) | |
| NYHA HF class I-II | 2139 (62.93 %) | 1108 (65.25 %) | |
| NYHA HF class III-IV | 1260 (37.07 %) | 590 (34.74 %) | |
| CCS Angina class 0-II * | 2551 (75.05 %) | 1322 (77.86 %) | |
| CCS Angina class III-IV * | 95 (2.79 %) | 42 (2.47 %) | |
| Atrial fibrillation | 1017 (29.92 %) | 493 (29.03 %) | |
| LVEDD in mm, mean (std) | 62.37 (10.32) | 62.71 (9.47) | |
| LVESD in mm, mean (std) | 52.04 (14.15) | 52.46 (15.30) | |
| Urgency - in hospital (vs. out of hospital) | 575 (16.92 %) | 252 (14.8 %) | |
| Non-Cardiac History | Hypertension | 1991 (58.58 %) | 966 (56.89 %) |
| Cerebrovascular disease/transient ischemic attack | 434 (12.77 %) | 236 (13.90 %) | |
| Peripheral vascular disease | 358 (10.53 %) | 181 (10.66 %) | |
| Chronic obstructive pulmonary disease | 556 (16.36 %) | 308 (18.14 %) | |
| Diabetes | 1298 (38.19 %) | 621 (36.57 %) | |
| Dialysis | 36 (1.06 %) | 16 (0.94 %) | |
| Current smoker * | 498 (14.65 %) | 246 (14.49 %) | |
| Ever smoker | 1981 (58.28 %) | 976 (57.48 %) | |
| Medications | Beta blockers | 2974 (87.50 %) | 1507 (88.75 %) |
| ACEI | 2531 (74.46 %) | 1249 (73.56 %) | |
| ARB | 622 (18.30 %) | 299 (17.61 %) | |
| ACEI or ARB | 3051 (89.76 %) | 1504 (88.57 %) | |
| Amiodarone | 318 (9.36 %) | 174 (10.25 %) | |
| Other class III antiarrhythmic | 12 (0.35 %) | <6 | |
| Digoxin | 862 (25.36 %) | 401 (23.62 %) | |
| Oral anticoagulation | 1081 (31.80 %) | 566 (33.33 %) | |
| Aspirin | 2097 (61.69 %) | 994 (58.54 %) | |
| Clopidogrel | 620 (18.24 %) | 309 (18.20 %) | |
| Loop diuretic | 2204 (64.84 %) | 1101 (64.84 %) | |
| Biomarkers | Sodium, mean (std) | 138.42 (3.59) | 138.64 (3.36) |
| Creatinine, mean (std) | 110.84 (59.92) | 108.75 (54.93) | |
| BUN, mean (std) | 9.37 (8.96) | 9.43 (9.50) | |
| eGFR, mean (std) | 68.82 (29.97) | 68.97 (32.42) | |
| Hemoglobin, mean (std) | 135.18 (16.98) | 134.77 (17.20) | |
| QRS_Manual, mean (std) | 130.29 (35.40) | 129.09 (35.81) | |
| QRS_Computer, mean (std) | 134.32 (35.06) | 133.65 (36.04) | |
| Device Type | Single chamber | 1442 (42.42 %) | 712 (41.93 %) |
| Dual chamber | 782 (23.01 %) | 398 (23.44 %) | |
| CRT-D | 1173 (34.51 %) | 587 (34.57 %) | |
| Manufacturer | Medtronic | 1659 (48.81 %) | 856 (50.41 %) |
| St. Jude Medical | 997 (29.33 %) | 489 (28.80 %) | |
| Boston Scientific | 665 (19.56 %) | 306 (18.02 % | |
| Sorin | 69 (2.03 %) | 46 (2.71 %) | |
| Biotronik | 9 (0.26 %) | <6 | |
| Lead Characteristics | Dual coil RV lead | 2971 (87.43 %) | 1477 (86.98 %) |
| Single coil RV lead | 423 (12.45 %) | 216 (12.72 %) | |
| Active RV fixation | 3192 (93.94 %) | 1594 (93.88 %) | |
| Passive fixation | 202 (5.94 %) | 100 (5.89 %) | |
| Institution | Hamilton | 599 (17.62 %) | 287 (16.9 %) |
| London | 502 (14.77 %) | 233 (13.72 %) | |
| Greater Toronto Area | 2040 (49.07 %) | 824 (48.53 %) | |
| Kingston | 258 (7.59 %) | 157 (9.25 %) | |
| Ottawa | 372 (10.94 %) | 197 (11.6 %) | |
| Provider Characteristics | Canadian Medical Graduate | 2196 (68.71 %) | 1084 (67.58 %) |
| Main specialty of Cardiology | 2827 (88.43 %) | 1417 (88.34 %) | |
| Volume of ICD inserted in last fiscal year, mean (std) | 124.15 (120.69) | 123.32 (120.79) | |
| Years since medical school graduation, mean (std) | 19.32 (7.94) | 19.06 (7.70) |
3.2. Outcome events
Patients were followed for 10-years after device implant, during which time 1832 deaths occurred in the derivation cohort. Cumulative death rate was 5.5 % (n = 186), 16.4 % (n = 555), 28.1 % (n = 952), 44.7 % (n = 1510) and 54.3 % (n = 1846) at 1-, 3- 5-, 8-, and 10-years respectively. Mortality data were available for all patients. A Kaplan Meier survival curve is shown in Fig. 2.
Fig. 2.
Survival after ICD insertion for primary prevention.
3.3. Multivariate regression models for the prediction of death
Multivariate predictors of time to death were modelled using a Cox regression model. After the initial backward selection, no socioeconomic or device factors remained significant. A simplified prediction model using only age, sex, and disease etiology had good calibration and moderate discrimination with an AUC of 0.70 (95 % CI 0.68–0.72) and 0.71 (0.68, 0.73) in the derivation and internal validation cohorts respectively. The stepwise addition of disease factors, medications, and biomarkers continued to improve model discrimination (Table 2). However, model calibration deteriorated after the addition of medications. Cubic spline testing using 5 knots suggested that the biomarkers sodium, renal function, hemoglobin, and QRS duration may not have a linear relationship with the outcome, so these variables were subsequently categorized using clinical significance and cubic spline thresholds. Categorized biomarkers improved model discrimination and calibration, with the calibration improving further with the removal of patient medications. The final model parameters are shown in Table 3 showing good discrimination with AUC 0.79 (0.77, 0.81) and 0.78 (0.76, 0.79) in derivation and internal validation cohorts respectively. This model also showed good calibration when assessed visually (Fig. 3). Increasing age, procedure urgency, symptoms assessed using NYHA HF scale, LVEF <20 %, comorbidities, hyponatremia, anemia, renal dysfunction, and prolonged QRS duration were associated with increased risk of death. Female sex and nonischemic cardiomyopathy were associated with decreased risk of death. Missing predictor data was relatively low at 7.9 % and 7.1 % in the derivation and validation cohorts respectively.
Table 2.
Model performance assessed using Akaike information criterion and area under the curve for the development of the base model.
Abbreviations: ECG = electrocardiogram, LVEF = left ventricular ejection fraction, NYHA = New York Heart Association dyspnea scale.
| Models | Predictors | Derivation Cohort IPCW AUC (95 % CI) | Validation Cohort IPCW AUC (95 % CI) |
|---|---|---|---|
| Model 1 | Age, sex, etiology | 0.70 (0.68, 0.72) | 0.71 (0.68, 0.73) |
| Model 2 | Age, sex, etiology, LVEF, NYHA | 0.74 (0.72, 0.75) | 0.73 (0.70, 0.75) |
| Model 3 | Age, sex, etiology, LVEF, NYHA, comorbiditiesa | 0.77 (0.75, 0.79) | 0.76 (0.74, 0.78) |
| Model 4 | Age, sex, etiology, LVEF, NYHA, comorbiditiesa, medicationsb | 0.78 (0.77, 0.80) | 0.77 (0.75, 0.79) |
| Model 5 | Age, sex, etiology, LVEF, NYHA, comorbiditiesa, medicationsb, biomarkersc | 0.79 (0.78, 0.81) | 0.78 (0.45, 1) |
| Model 6a | Age, sex, etiology, LVEF, NYHA, comorbiditiesa, medicationsb, categorized biomarkersd | 0.80 (0.78, 0.81) | 0.78 (0.76, 0.80) |
| Model 6b | Age, sex, etiology, LVEF, NYHA, comorbiditiesa, categorized biomarkersd | 0.79 (0.77, 0.81) | 0.78 (0.76, 0.79) |
Comorbidities: diabetes, hypertension, cerebrovascular accident/transient ischemic attack, current smoker, peripheral vascular disease, chronic obstructive pulmonary disease, dialysis.
Medications: loop diuretic, oral anticoagulant, digoxin, aspirin, clopidogrel.
Biomarkers: sodium, hemoglobin, estimated glomerular filtration, QRS on ECG.
Categories for biomarkers: Sodium:≤135 mmol/L, 136–140 mmol/L, >140 mmol/L; Hemoglobin:≤70 g/L, 71–120 g/L, 121–140 g/L, >140 g/L; eGFR:≤60 mL/min/1.73 m2, >60 mL/min/1.73 m2; QRS:≤120 ms, 121–160 ms, >160 ms
Table 3.
Predictors of death after ICD implant for primary prevention
Abbreviations: CM = cardiomyopathy; COPD = chronic obstructive pulmonary disease; CVA = cerebrovascular accident; eGFR = estimated glomerular filtration rate; HF = heart failure; NYHA = New York Heart Association; TIA = transient ischemic attack.
| Parameter | Category | Parameter Estimate | Hazard Ratio | 95 % CI | p-value |
|---|---|---|---|---|---|
| Sex | female | −0.20141 | 0.82 | 0.72, 0.93 | 0.0029 |
| Age (years) | … | 0.04129 | 1.04 | 1.04, 1.05 | <.0001 |
| Etiology (reference = ischemic CM) | Non-ischemic | −0.10944 | 0.90 | 0.79, 1.01 | 0.1571 |
| Other | −0.25612 | 0.77 | 0.42, 1.41 | ||
| Urgency (reference = ambulatory) | In hospital | 0.12252 | 1.13 | 0.99, 1.28 | 0.0602 |
| NYHA HF Scale (reference = I-II) | III-IV | 0.38306 | 1.47 | 1.32, 1.63 | <.0001 |
| Left ventricular ejection fraction (reference = >35 %) | 0–20 % | 0.25348 | 1.29 | 1.05, 1.58 | 0.0001 |
| 21–30 % | 0.00722 | 1.01 | 0.83, 1.22 | ||
| 31–35 % | −0.05632 | 0.95 | 0.75, 1.18 | ||
| Diabetes | … | 0.34677 | 1.41 | 1.28, 1.57 | <.0001 |
| Hypertension | … | −0.07502 | 0.93 | 0.84, 1.03 | 0.1568 |
| CVA/TIA | … | 0.19752 | 1.22 | 1.06, 1.40 | 0.0043 |
| Current smoker | … | 0.32417 | 1.38 | 1.20, 1.60 | <.0001 |
| Peripheral vascular disease | … | 0.35176 | 1.42 | 1.23, 1.64 | <.0001 |
| COPD | … | 0.36865 | 1.45 | 1.28, 1.63 | <.0001 |
| Dialysis | … | 0.80191 | 2.23 | 1.51, 3.30 | <.0001 |
| Sodium (reference = 135–140 mmol/L) | ≤135 mmol/L | 0.30327 | 3.31 | 0.82, 13.40 | <.0001 |
| >140 mmol/L | 0.02762 | 1.32 | 1.16, 1.50 | ||
| Hemoglobin (reference = 120–140 g/L) | ≤70 g/L | 1.19812 | 0.82 | 0.73, 0.92 | <.0001 |
| 70–120 g/L | 0.27725 | 1.35 | 1.19, 1.54 | ||
| >140 g/L | −0.19455 | 1.03 | 0.91, 1.16 | ||
| eGFR | ≤60 mL/min/1.73 m2 | 0.46828 | 1.60 | 1.44, 1.78 | <.0001 |
| QRS (reference = <120 ms) | 120–160 ms | 0.12840 | 1.14 | 1.01, 1.28 | 0.0405 |
| >160 ms | −0.00661 | 0.99 | 0.87, 1.13 |
Fig. 3.
Model calibration using observed vs predicted risk of death for the development of the base model. Model name, predictors included in each model, and corresponding derivation and validation calibration plots are indicated in columns a to d respectively. Diagonal dotted line indicates perfect calibration. Seven models were sequentially tested, Model 7 was identified as the final model with the best performance in calibration and discrimination. Abbreviations: ECG = electrocardiogram, LVEF = left ventricular ejection fraction, NYHA = New York Heart Association dyspnea scale 1Comorbidities: diabetes, hypertension, cerebrovascular accident/transient ischemic attack, current smoker, peripheral vascular disease, chronic obstructive pulmonary disease, dialysis 2Medications: loop diuretic, oral anticoagulant, digoxin, aspirin, clopidogrel 3Biomarkers: sodium, hemoglobin, estimated glomerular filtration, QRS on ECG 4Categories for biomarkers: Sodium:≤135 mmol/L, 136–140 mmol/L, >140 mmol/L; Hemoglobin:≤70 g/L, 71–120 g/L, 121–140 g/L, >140 g/L; eGFR:≤60 mL/min/1.73 m2, >60 mL/min/1.73 m2; QRS:≤120 ms, 121–160 ms, >160 ms
3.4. Sensitivity analyses
Several sensitivity analyses were performed to ensure that the best prediction model was chosen within data limitations. Sensitivity analyses investigated the effect of additional variables of unclear significance including device type, provider factors (provider main specialty, Canadian medical graduate status), implantation site, frailty score, and comorbidity burden using the Johns Hopkins ACG comorbidity index. Since left ventricular end diastolic dimension (LVEDD) and implanter volume may be important variables which had significant missing variables, they were included in a sensitivity analysis after using missing variable imputation. Sex-etiology interaction terms and allowing for maximal follow-up of up to 14 years without censoring were also studied. Sensitivity model performance was assessed in the derivation and internal validation cohorts, and discrimination and calibration are reported in Supplement S1 and S2 respectively. None of the sensitivity analyses showed improved performance when compared to the final base model.
4. Discussion
As ICD implant rates continue to rise and healthcare providers adopt the use of CRT devices [22,23], it can often be challenging to translate randomized controlled trial evidence to clinical practice. Follow-up of ICD recipients in clinical practice indicate that complications and morbidity after ICD implant remain high [[24], [25], [26]]. Since the average lifespan of ICD devices ranges from 5- to 10-years [27], it is crucial to have accurate long-term survival data in order to have well-informed patient-centered discussions when assessing patients for ICD therapy.
To our knowledge, our study is the first prognostic model using population-based data to predict 10-year survival using baseline patient and demographic predictors determined at the time of evaluation for ICD insertion. Since Ontario utilizes a universal single-payer healthcare system where all devices are paid for by the Ministry of Health and Long-term Care, and data was collected at ICES without requirement of informed consent, we were uniquely able to capture accurate long-term patient outcomes in clinical practice on a population-level.
4.1. Survival
Only a few studies have reported long-term survival rates of patients after receiving ICD therapy for primary prevention. Follow-up of the MADIT II population found a median survival of 8 years, with cumulative probability of death of 40 % at 6 years [28] and 49 % at 8 years [29]. Long-term follow-up of the SCD-HeFT population found a 52.5 % death rate at 10-years [30]. Cumulative death rates in our study using the Ontario ICD database were similar to previous observational studies [14,15,31,32], with 1, 3, 5, 8, and 10-year cumulative death rates of 5.5 % (n = 186), 16.4 % (n = 555), 28.1 % (n = 952), 44.7 % (n = 1510), and 54.3 % (n = 1846) respectively. Survival at 1- and 3-years predicted using this model performed similarly in terms of discrimination and calibration to previously published models with AUC 0.73 (95 % CI 0.68, 0.79) and AUC 0.71 (95 % CI 0.67, 0.74) respectively in the internal validation cohort [33].
4.2. Predictors of survival
Increased age has been identified as a strong predictor of death after ICD in several observational trials [[13], [15], [28], [34], [35]], [[,15,28,34,35] as well as higher risk of complication after device insertion [36]. It is postulated that with advancing age, patients are at increasing risk of death by non-arrhythmic causes [37]. There may be decreased ICD benefit with increased age as evidenced in post-hoc analysis of the DANISH trial [38]. Interestingly, previous research from using the Ontario ICD database has shown that there is no significant decline in the rate of appropriate ICD shocks when using competing risk analysis [35]. However, the relative morbidity associated with sequelae of appropriate ICD therapy for older adults vs. younger adults is unknown. It is plausible that while older adults continue to receive appropriate shocks, they may have increased emergency department visits and longer hospitalizations which are subsequently translated to increased morbidity and mortality when compared to younger populations. Our analysis helps to further highlight that increasing age is a strong predictor of death even in a multivariate model controlling for several patient comorbidities and biomarkers up to 10-years after ICD implant.
Female patients were significantly under-represented in the initial landmark ICD trials, with most enrolling 75–90 % male patients [[4], [5], [6]]. While ICD therapy is assumed to be equally beneficial in women and men, recent studies have identified that women were less likely to receive appropriate ICD therapy via anti-tachycardia therapy or ICD shocks [[39], [40], [41]] and more likely to experience complications after device implantation [36,39,42]. However, women are more likely to benefit from CRT [43]. This may be due to less sudden death in women when adjusted for heart failure severity [44]. Our multi-center analysis showed higher survival rates in women with primary prevention ICD therapy when compared to men persisting up to 10-years after insertion, consistent with previous observational studies [26,40,42]. However, two meta-analyses of clinical trials did not find a significant reduction of sudden cardiac death in women implanted with ICD for primary prevention [45,46]. Further studies in women are required to better understand the benefit of ICD therapy in this population.
Similar predictors of death were identified by the BaSIS model, which weighed risk of appropriate shock against 1-year risk of death after ICD insertion using data from the Ontario ICD Database [47]. In both models, increasing age, ischemic cardiomyopathy, worse heart failure symptoms, diabetes, smoking history, COPD, impaired renal function, hyponatremia, and anemia were identified as predictors of increased risk of death. Prior revascularization with percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) was associated with improved survival at 1-year in the BaSIS model but were not significant predictors of outcomes in our model. Additionally, we found that diabetes, prior CVA/stroke, peripheral vascular disease, and dialysis were also associated with increased risk of death, with improved model performance when medications were excluded from the final prediction model. This is likely because any additional predictive power afforded with the addition of medications was counteracted by model over-fitting and collinearity between medications and other predictor variables (e.g. comorbidities).
4.3. Sensitivity analyses
Several additional models were constructed during sensitivity analyses and compared to the final model (Supplement S1-2). Previous literature suggests that increasing device complexity is associated with increased complications and short-term risk of death, with dual-chamber and CRT devices having more complications than single-chamber devices [[15], [36], [48]], [[,36,48] though it was not associated with death in our model. Addition of device type did not improve model performance.
We also assessed the effect of adding provider factors including Canadian Medical Graduate status, physician main specialty, and implantation center to the final model. While provider factors are not classically included in prognostic models, our goal was to assess if provider factors played a role in changing patient outcomes. We did not find that the addition of provider factors improved our predictive model.
Previous prognostic scores developed to predict outcomes after ICD insertion lack measurement of patient frailty, which can be an important decision factor for clinicians when discussing ICD therapy with patients. We attempted to address this with the inclusion of the Hospital Frailty Risk Score and the Johns Hopkins ACG comorbidity index, which identify patients at high risk of prolonged hospital admission and comorbidity burden respectively [18,49] using health services data. Addition of these surrogate measures of frailty did not affect model performance.
While missing predictor data was not a significant issue in our model, two potentially important predictors were excluded from the original model due to missing data. Enlarged left ventricular dimension [36] and low implanter volume [24] have previously been associated with increased risk of complications and subsequent mortality in previous research from the Ontario ICD database. Our analysis showed that after single imputation of missing variables, the addition of left ventricular end diastolic dimension and implanter volume in the last fiscal year did not improve model performance.
Previous observational studies have identified that women receiving ICD therapy are more likely to have NICM when compared to men. To further investigate whether sex may modify the ability of disease factors to predict long-term outcomes after ICD insertion, several interaction terms between sex and disease etiology, NYHA HF symptoms, and LVEF were explored. After initial multivariate analysis, only the sex and disease etiology interaction term remained significant (p = 0.04). Addition of the sex and disease etiology interaction term did not improve model performance.
Finally, allowing for maximal follow-up up to 14 years after ICD insertion also did not improve model performance.
4.4. Strengths and limitations
There are some notable limitations to our study. First, while the Ontario ICD database is broadly inclusive of patients receiving ICD therapy in Ontario, recipients must have an active Ontario Health Card. Generalization to military populations as well as populations outside of Ontario are unknown. Second, our model was not validated using an independent external database. While external validation is crucial prior to using prognostic models in clinical practice, our research helps to conceptualize the import predictors of long-term outcomes in patients receiving ICD therapy. There is a potential risk of misclassification and loss of information with the categorization of non-linear predictor variables (LVEF, serum hemoglobin, sodium, QRS interval). While polynomial transformations and cubic spline can be used to model non-linearity, the use of these methods would further complicate a clinical risk score meant to be used at the bedside. While we attempted to capture the most important baseline variables which may affect long-term outcomes, there may still be unmeasured confounders such as history of malignancy and chemotherapy. Finally, recent advances in goal-directed medical therapy (GDMT) for medical management of heart failure may modify the population of patients referred for primary prevention ICD. While we expect similar factors to be important in the prediction of long-term outcomes, our model would need to be re-validated in a long-term follow-up of contemporary heart failure patients. These limitations are outweighed by the unique strengths of this study, including the completeness of data on patient outcomes in long-term follow-up, the population level data collection with minimal risk of selection bias, and ability to predict 10-year outcomes using only baseline clinical information collected at time of initial ICD assessment.
4.5. Clinical significance and future directions
We were able to identify a new prognostic model using baseline clinical predictors to predict long-term outcomes after ICD insertion for primary prevention. Using several additional sensitivity analyses, we were able to find the simplest model with the best statistical performance to identify important predictor variables (Table 3) using internal validation. Risk stratifying patients into those who are at high and low risk of early death can inform clinicians and patients in shared decision making in determining when ICD therapy is most appropriate.
There are several areas of additional research which can help identify the best care of patients with primary prevention ICD therapy beyond baseline patient characteristics. Competing risk of dying from non-arrhythmic causes, risk of device implantation related complications, and the impact of ICD shocks are important considerations when choosing patients for ICD insertion for primary prevention. These would be better studied using a competing risks prediction model with time-dependent predictors such as device infections, ICD shocks, and device revisions.
Finally, previous literature has identified that despite the mortality benefit obtained with ICD therapy for primary prevention, healthcare utilization in this patient population remains high. The ability to predict patterns of healthcare utilization at baseline as well as studies into costs of care of patients after ICD implant would be invaluable for health policy makers to optimally allocate healthcare resources. The Ontario ICD database would be uniquely able to study this relationship due to data linkage with other acute and ambulatory health-care resource utilization databases housed within ICES.
Conclusion
In this paper, we present a new prognostic model which accurately assesses prognosis up to 10-years after ICD insertion for primary prevention. The model uses baseline clinical predictors easily available to clinicians at the time of evaluation for ICD therapy and performed well with good discrimination and calibration in an internal validation cohort. Accurate personalized prediction of outcomes after ICD insertion can be used in shared decision making when counselling patients for ICD therapy.
Ethics statement
Approval was obtained from Queen's University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board (HSREB). Department Code; SMED-249-21, TRAQ # 6034674.
Sources of funding
This study was supported by a Foundation grant from the Canadian Institutes of Health Research (grant # FDN 148446). Dr. Lee is the Ted Rogers Chair in Heart Function Outcomes, University Health Network, University of Toronto.
Data availability statement
The data associated with this study have not been deposited into a publicly available repository, as the data are confidential.
CRediT authorship contribution statement
Chang Nancy Wang: Writing - review & editing, Writing - original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Zihang Lu: Writing - review & editing, Validation, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Christopher S. Simpson: Visualization, Validation, Supervision, Conceptualization. Douglas S. Lee: Writing - review & editing, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Joan E. Tranmer: Writing - review & editing, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Douglas Lee reports financial support was provided by Canadian Institutes of Health Research. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This study was supported by the ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long‐Term Care (MLTC). Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information and Ontario MOH. The analyses, conclusions, opinions and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. This document used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from ©Canada Post Corporation and Statistics Canada. We thank the Toronto Community Health Profiles Partnership for providing access to the Ontario Marginalization Index.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.heliyon.2023.e23355.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
References
- 1.Al-Khatib S.M., Stevenson W.G., Ackerman M.J., Bryant W.J., Callans D.J., Curtis A.B., Deal B.J., Dickfeld T., Field M.E., Fonarow G.C., Gillis A.M., Granger C.B., Hammill S.C., Hlatky M.A., Joglar J.A., Neal K.G., Matlock D.D., Myerburg R.J., Page R.L. AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Circulation. 2017;138(13):e272–e391. doi: 10.1161/CIR.0000000000000549. 2018. [DOI] [PubMed] [Google Scholar]
- 2.Ezekowitz J.A., O'Meara E., McDonald M.A., Abrams H., Chan M., Ducharme A., Giannetti N., Grzeslo A., Hamilton P.G., Heckman G.A., Howlett J.G., Koshman S.L., Lepage S., McKelvie R.S., Moe G.W., Rajda M., Swiggum E., Virani S.A., Zieroth S., Al-Hesayen A., Cohen-Solal A., D'Astous M., De S., Estrella-Holder E., Fremes S., Green L., Haddad H., Harkness K., Hernandez A.F., Kouz S., LeBlanc M.-H., Masoudi F.A., Ross H.J., Roussin A., Sussex B. 2017 comprehensive update of the Canadian cardiovascular society guidelines for the management of heart failure. Can. J. Cardiol. 2017;33(11):1342–1433. doi: 10.1016/j.cjca.2017.08.022. [DOI] [PubMed] [Google Scholar]
- 3.Priori S.G., Blomström-Lundqvist C., Mazzanti A., Blom N., Borggrefe M., Camm J., Elliott P.M., Fitzsimons D., Hatala R., Hindricks G., Kirchhof P., Kjeldsen K., Kuck K.-H., Hernandez-Madrid A., Nikolaou N., Norekvål T.M., Spaulding C., Van Veldhuisen D.J., Kolh P., Lip G.Y.H., Agewall S., Barón-Esquivias G., Boriani G., Budts W., Bueno H., Capodanno D., Carerj S., Crespo-Leiro M.G., Czerny M., Deaton C., Dobrev D., Erol Ç., Galderisi M., Gorenek B., Kriebel T., Lambiase P., Lancellotti P., Lane D.A., Lang I., Manolis A.J., Morais J., Moreno J., Piepoli M.F., Rutten F.H., Sredniawa B., Zamorano J.L., Zannad F., Zamorano J.L., Aboyans V., Achenbach S., Agewall S., Badimon L., Barón-Esquivias G., Baumgartner H., Bax J.J., Bueno H., Carerj S., Dean V., Erol Ç., Fitzsimons D., Gaemperli O., Kirchhof P., Kolh P., Lancellotti P., Lip G.Y., Nihoyannopoulos P., Piepoli M.F., Ponikowski P., Roffi M., Torbicki A., Vaz Carneiro A., Windecker S., Piruzyan A., Roithinger F.X., Mairesse G.H., Goronja B., Shalganov T., Puljevic D., Antoniades L., Kautzner J., Larsen J.M., Aboulmaaty M., Kampus P., Hedman A., Kamcevska-Dobrkovic L., Piot O., Etsadashvili K., Eckardt L., Deftereos S., Gellér L., Gizurarson S., Keane D., Haim M., Bella P.D., Abdrakhmanov A., Mirrakhimov A., Kalejs O., Lamin H.B., Marinskis G., Groben L., Sammut M., Raducan A., Chaib A., Tande P.M., Lenarczyk R., Morgado F.B., Vatasescu R., Mikhaylov E.N., Hlivak P., Arenal A., Jensen-Urstad M., Sticherling C., Zeppenfeld K., Chettaoui R., Demir M., Duncan E., Parkhomenko A. 2015 ESC guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac deathThe task force for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death of the European society of cardiology (ESC)endorsed by: association for European paediatric and congenital cardiology (AEPC) Eur. Heart J. 2015;36(41):2793–2867. doi: 10.1093/eurheartj/ehv316. [DOI] [PubMed] [Google Scholar]
- 4.Buxton A.E., Lee K.L., Fisher J.D., Josephson M.E., Prystowsky E.N., Hafley G. A randomized study of the prevention of sudden death in patients with coronary artery disease. Multicenter Unsustained Tachycardia Trial Investigators. N. Engl. J. Med. 1999;341(25):1882–1890. doi: 10.1056/NEJM199912163412503. [DOI] [PubMed] [Google Scholar]
- 5.Moss A.J., Zareba W., Hall W.J., Klein H., Wilber D.J., Cannom D.S., Daubert J.P., Higgins S.L., Brown M.W., Andrews M.L., M.A.D.I.T.I. Investigators, prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced ejection fraction. N. Engl. J. Med. 2002;346(12):877–883. doi: 10.1056/NEJMoa013474. [DOI] [PubMed] [Google Scholar]
- 6.Bardy G.H., Lee K.L., Mark D.B., Poole J.E., Packer D.L., Boineau R., Domanski M., Troutman C., Anderson J., Johnson G., McNulty S.E., Clapp-Channing N., Davidson-Ray L.D., Fraulo E.S., Fishbein D.P., Luceri R.M., Ip J.H., S.C.D.i.H.F.T.S.-H. Investigators, amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N. Engl. J. Med. 2005;352(3):225–237. doi: 10.1056/NEJMoa043399. [DOI] [PubMed] [Google Scholar]
- 7.Connolly S.J., Gent M., Roberts R.S., Dorian P., Roy D., Sheldon R.S., Mitchell L.B., Green M.S., Klein G.J., O'Brien B. Canadian implantable defibrillator study (CIDS) Circulation. 2000;101(11):1297–1302. doi: 10.1161/01.cir.101.11.1297. [DOI] [PubMed] [Google Scholar]
- 8.Kadish A., Dyer A., Daubert J.P., Quigg R., Estes N.A.M., Anderson K.P., Calkins H., Hoch D., Goldberger J., Shalaby A., Sanders W.E., Schaechter A., Levine J.H., D.i.N.-I.C.T.E.D. Investigators, prophylactic defibrillator implantation in patients with nonischemic dilated cardiomyopathy. N. Engl. J. Med. 2004;350(21):2151–2158. doi: 10.1056/NEJMoa033088. [DOI] [PubMed] [Google Scholar]
- 9.Køber L., Thune J.J., Nielsen J.C., Haarbo J., Videbæk L., Korup E., Jensen G., Hildebrandt P., Steffensen F.H., Bruun N.E., Eiskjær H., Brandes A., Thøgersen A.M., Gustafsson F., Egstrup K., Videbæk R., Hassager C., Svendsen J.H., Høfsten D.E., Torp-Pedersen C., Pehrson S., Investigators D. Defibrillator implantation in patients with nonischemic systolic heart failure. N. Engl. J. Med. 2016;375(13):1221–1230. doi: 10.1056/NEJMoa1608029. [DOI] [PubMed] [Google Scholar]
- 10.Moss A.J., Hall W.J., Cannom D.S., Klein H., Brown M.W., Daubert J.P., Estes N.A.M., Foster E., Greenberg H., Higgins S.L., Pfeffer M.A., Solomon S.D., Wilber D., Zareba W., Investigators M.-C.T. Cardiac-resynchronization therapy for the prevention of heart-failure events. N. Engl. J. Med. 2009;361(14):1329–1338. doi: 10.1056/NEJMoa0906431. [DOI] [PubMed] [Google Scholar]
- 11.Al-Khatib S.M., Hellkamp A., Bardy G.H., Hammill S., Hall W.J., Mark D.B., Anstrom K.J., Curtis J., Al-Khalidi H., Curtis L.H., Heidenreich P., Peterson E.D., Sanders G., Clapp-Channing N., Lee K.L., Moss A.J. Survival of patients receiving a primary prevention implantable cardioverter-defibrillator in clinical practice vs clinical trials. JAMA. 2013;309(1):55–62. doi: 10.1001/jama.2012.157182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dickstein K., Normand C., Auricchio A., Bogale N., Cleland J.G., Gitt A.K., Stellbrink C., Anker S.D., Filippatos G., Gasparini M., Hindricks G., Lundqvist C.B., Ponikowski P., Ruschitzka F., Botto G.L., Bulava A., Duray G., Israel C., Leclercq C., Margitfalvi P., Cano Ó., Plummer C., Sarigul N.U., Sterlinski M., Linde C. CRT Survey II: a European Society of Cardiology survey of cardiac resynchronisation therapy in 11 088 patients—who is doing what to whom and how? Eur. J. Heart Fail. 2018;20(6):1039–1051. doi: 10.1002/ejhf.1142. [DOI] [PubMed] [Google Scholar]
- 13.Agarwal S.K., Singla I., Hreybe H., Saba S. Predictors of mortality amongst recipients of implantable cardioverter defibrillators for secondary prevention of cardiac arrest. Indian Pacing Electrophysiol J. 2007;7(4):218–224. [PMC free article] [PubMed] [Google Scholar]
- 14.Ranasinghe I., Parzynski C.S., Freeman J.V., Dreyer R.P., Ross J.S., Akar J.G., Krumholz H.M., Curtis J.P. Long-term risk for device-related complications and reoperations after implantable cardioverter-defibrillator implantation. Ann. Intern. Med. 2016;165(1):20–29. doi: 10.7326/M15-2732. [DOI] [PubMed] [Google Scholar]
- 15.Hawkins N.M., Grubisic M., Andrade J.G., Huang F., Ding L., Gao M., Bashir J. Long-term complications, reoperations and survival following cardioverter-defibrillator implant. Heart (British Cardiac Society) 2018;104(3):237–243. doi: 10.1136/heartjnl-2017-311638. [DOI] [PubMed] [Google Scholar]
- 16.Lee D.S., Birnie D., Cameron D., Crystal E., Dorian P., Gula L.J., Healey J.S., Janmohammed A., Khaykin Y., Krahn A.D., LeFeuvre C., Simpson C.S., Yee R., Hardy J., Slaughter P.M., Chen Z., Alter D.A., Laupacis A., Tu J.V. Design and implementation of a population-based registry of implantable cardioverter defibrillators (ICDs) in Ontario. Heart Rhythm. 2008;5(9):1250–1256. doi: 10.1016/j.hrthm.2008.05.015. [DOI] [PubMed] [Google Scholar]
- 17.Tu J.V., Willison D.J., Silver F.L., Fang J., Richards J.A., Laupacis A., Kapral M.K. 2004. Impracticability of Informed Consent in the Registry of the Canadian Stroke Network. [DOI] [PubMed] [Google Scholar]
- 18.Gilbert T., Neuburger J., Kraindler J., Keeble E., Smith P., Ariti C., Arora S., Street A., Parker S., Roberts H.C., Bardsley M., Conroy S. Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391(10132):1775–1782. doi: 10.1016/S0140-6736(18)30668-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Booth C.M., Li G., Zhang-Salomons J., Mackillop W.J. The impact of socioeconomic status on stage of cancer at diagnosis and survival. Cancer. 2010;116(17):4160–4167. doi: 10.1002/cncr.25427. [DOI] [PubMed] [Google Scholar]
- 20.Royston P., Moons K.G.M., Altman D.G., Vergouwe Y. Prognosis and prognostic research: developing a prognostic model. BMJ. 2009;338:b604. doi: 10.1136/bmj.b604. [DOI] [PubMed] [Google Scholar]
- 21.N.R. Cook, SAS Macros - Risk Prediction Modeling Division of Preventive Medicine. Hosmer-Lemeshow Goodness of Fit Test. https://ncook.bwh.harvard.edu/sas-macros.html.
- 22.Bottle A., Faitna P., Aylin P., Cowie M.R. Five-year survival and use of hospital services following ICD and CRT implantation: comparing real-world data with RCTs. ESC Heart Failure. 2021;8(4):2438–2447. doi: 10.1002/ehf2.13357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Boveda S., Narayanan K., Jacob S., Providencia R., Algalarrondo V., Bouzeman A., Beganton F., Defaye P., Perier M.-C., Sadoul N., Piot O., Klug D., Gras D., Fauchier L., Bordachar P., Babuty D., Deharo J.-C., Leclercq C., Marijon E. Temporal trends over a decade of defibrillator therapy for primary prevention in community practice. J. Cardiovasc. Electrophysiol. 2017;28(6):666–673. doi: 10.1111/jce.13198. [DOI] [PubMed] [Google Scholar]
- 24.Krahn A.D., Lee D.S., Birnie D., Healey J.S., Crystal E., Dorian P., Simpson C.S., Khaykin Y., Cameron D., Janmohamed A., Yee R., Austin P.C., Chen Z., Hardy J., Tu J.V. Predictors of short-term complications after implantable cardioverter-defibrillator replacement. Circulation: Arrhythmia Electrophysiol. 2011;4(2):136–142. doi: 10.1161/CIRCEP.110.959791. [DOI] [PubMed] [Google Scholar]
- 25.Trivedi A.N., Sequist T.D., Ayanian J.Z. 2006. Impact of Hospital Volume on Racial Disparities in Cardiovascular Procedure Mortality. [DOI] [PubMed] [Google Scholar]
- 26.Heijden A.C.v.d., Thijssen J., Borleffs C.J.W., Rees J.B.v., Höke U., Velde E.T.v.d., Erven L.v., Schalij M.J. Gender-specific differences in clinical outcome of primary prevention implantable cardioverter defibrillator recipients. Heart. 2013;99(17):1244–1249. doi: 10.1136/heartjnl-2013-304013. [DOI] [PubMed] [Google Scholar]
- 27.Zanon F., Martignani C., Ammendola E., Menardi E., Narducci M.L., Filippo P.D., Santamaria M., Campana A., Stabile G., Potenza D.R., Pastore G., Iori M., Rosa C.L., Biffi M. Device longevity in a contemporary cohort of ICD/CRT-D patients undergoing device replacement. J. Cardiovasc. Electrophysiol. 2016;27(7):840–845. doi: 10.1111/jce.12990. [DOI] [PubMed] [Google Scholar]
- 28.Cygankiewicz I., Gillespie J., Zareba W., Brown M.W., Goldenberg I., Klein H., McNitt S., Polonsky S., Andrews M., Dwyer E.M., Hall W.J., Moss A.J. Predictors of long-term mortality in multicenter automatic defibrillator implantation trial II (MADIT II) patients with implantable cardioverter-defibrillators. Heart Rhythm. 2009;6(4):468–473. doi: 10.1016/j.hrthm.2008.12.023. [DOI] [PubMed] [Google Scholar]
- 29.Ilan G., John G., M A J., Jackson H.W., Helmut K., Scott M., B M W., Iwona C., Wojciech Z., null n. Long-term benefit of primary prevention with an implantable cardioverter-defibrillator. Circulation. 2010;122(13):1265–1271. doi: 10.1161/CIRCULATIONAHA.110.940148. [DOI] [PubMed] [Google Scholar]
- 30.Poole J.E., Olshansky B., Mark D.B., Anderson J., Johnson G., Hellkamp A.S., Davidson-Ray L., Fishbein D.P., Boineau R.E., Anstrom K.J., Reinhall P.G., Packer D.L., Lee K.L., Bardy G.H. Long-term outcomes of implantable cardioverter-defibrillator therapy in the SCD-HeFT. J. Am. Coll. Cardiol. 2020;76(4):405–415. doi: 10.1016/j.jacc.2020.05.061. [DOI] [PubMed] [Google Scholar]
- 31.van Welsenes G.H., van Rees J.B., Borleffs C.J.W., Cannegieter S.C., Bax J.J., van Erven L., Schalij M.J. Long-term follow-up of primary and secondary prevention implantable cardioverter defibrillator patients. EP Europace. 2011;13(3):389–394. doi: 10.1093/europace/euq494. [DOI] [PubMed] [Google Scholar]
- 32.Kramer D.B., Friedman P.A., Kallinen L.M., Morrison T.B., Crusan D.J., Hodge D.O., Reynolds M.R., Hauser R.G. Development and validation of a risk score to predict early mortality in recipients of implantable cardioverter-defibrillators. Heart Rhythm. 2012;9(1):42–46. doi: 10.1016/j.hrthm.2011.08.031. [DOI] [PubMed] [Google Scholar]
- 33.Alba A.C., Walter S.D., Guyatt G.H., Levy W.C., Fang J., Ross H.J., Lee D.S. Predicting survival in patients with heart failure with an implantable cardioverter defibrillator: the heart failure meta-score. J. Card. Fail. 2018;24(11):735–745. doi: 10.1016/j.cardfail.2017.11.002. [DOI] [PubMed] [Google Scholar]
- 34.Fudim M., Carlisle M.A., Devaraj S., Ajam T., Ambrosy A.P., Pokorney S.D., Al‐Khatib S.M., Kamalesh M. One-year mortality after implantable cardioverter-defibrillator placement within the veterans affairs health system. Eur. J. Heart Fail. 2020;22(5):859–867. doi: 10.1002/ejhf.1755. [DOI] [PubMed] [Google Scholar]
- 35.Yung D., Birnie D., Dorian P., Healey J.S., Simpson C.S., Crystal E., Krahn A.D., Khaykin Y., Cameron D., Chen Z., Lee D.S. Survival after implantable cardioverter-defibrillator implantation in the elderly. Circulation. 2013;127(24):2383–2392. doi: 10.1161/CIRCULATIONAHA.113.001442. [DOI] [PubMed] [Google Scholar]
- 36.Lee D.S., Krahn A.D., Healey J.S., Birnie D., Crystal E., Dorian P., Simpson C.S., Khaykin Y., Cameron D., Janmohamed A., Yee R., Austin P.C., Chen Z., Hardy J., Tu J.V. Evaluation of early complications related to de novo cardioverter defibrillator implantation: insights from the Ontario ICD database. J. Am. Coll. Cardiol. 2010;55(8):774–782. doi: 10.1016/j.jacc.2009.11.029. [DOI] [PubMed] [Google Scholar]
- 37.Shadman R., Poole J.E., Dardas T.F., Mozaffarian D., Cleland J.G.F., Swedberg K., Maggioni A.P., Anand I.S., Carson P.E., Miller A.B., Levy W.C. A novel method to predict the proportional risk of sudden cardiac death in heart failure: derivation of the Seattle Proportional Risk Model. Heart Rhythm. 2015;12(10):2069–2077. doi: 10.1016/j.hrthm.2015.06.039. [DOI] [PubMed] [Google Scholar]
- 38.Elming M.B., Nielsen J.C., Haarbo J., Videbæk L., Korup E., Signorovitch J., Olesen L.L., Hildebrandt P., Steffensen F.H., Bruun N.E., Eiskjær H., Brandes A., Thøgersen A.M., Gustafsson F., Egstrup K., Videbæk R., Hassager C., Svendsen J.H., Høfsten D.E., Torp-Pedersen C., Pehrson S., Køber L., Thune J.J. Age and outcomes of primary prevention implantable cardioverter-defibrillators in patients with nonischemic systolic heart failure. Circulation. 2017;136(19):1772–1780. doi: 10.1161/CIRCULATIONAHA.117.028829. [DOI] [PubMed] [Google Scholar]
- 39.MacFadden D.R., Crystal E., Krahn A.D., Mangat I., Healey J.S., Dorian P., Birnie D., Simpson C.S., Khaykin Y., Pinter A., Nanthakumar K., Calzavara A.J., Austin P.C., Tu J.V., Lee D.S. Sex differences in implantable cardioverter-defibrillator outcomes: findings from a prospective defibrillator database. Ann. Intern. Med. 2012;156(3):195–203. doi: 10.7326/0003-4819-156-3-201202070-00007. [DOI] [PubMed] [Google Scholar]
- 40.Seegers J., Conen D., Jung K., Bergau L., Dorenkamp M., Lüthje L., Sohns C., Sossalla S.T., Fischer T.H., Hasenfuss G., Friede T., Zabel M. Sex difference in appropriate shocks but not mortality during long-term follow-up in patients with implantable cardioverter-defibrillators. EP Europace. 2016;18(8):1194–1202. doi: 10.1093/europace/euv361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Sticherling C., Arendacka B., Svendsen J.H., Wijers S., Friede T., Stockinger J., Dommasch M., Merkely B., Willems R., Lubinski A., Scharfe M., Braunschweig F., Svetlosak M., Zürn C.S., Huikuri H., Flevari P., Lund-Andersen C., Schaer B.A., Tuinenburg A.E., Bergau L., Schmidt G., Szeplaki G., Vandenberk B., Kowalczyk E., Eick C., Juntilla J., Conen D., Zabel M., Investigators E.-C.-I. Sex differences in outcomes of primary prevention implantable cardioverter-defibrillator therapy: combined registry data from eleven European countries. EP Europace. 2018;20(6):963–970. doi: 10.1093/europace/eux176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Jamerson D., Mcnitt S., Polonsky S., Zareba W., Moss A., Tompkins C. Early procedure-related adverse events by gender in MADIT-CRT. J. Cardiovasc. Electrophysiol. 2014;25(9):985–989. doi: 10.1111/jce.12438. [DOI] [PubMed] [Google Scholar]
- 43.Woods B., Hawkins N., Mealing S., Sutton A., Abraham W.T., Beshai J.F., Klein H., Sculpher M., Plummer C.J., Cowie M.R. Individual patient data network meta-analysis of mortality effects of implantable cardiac devices. Heart. 2015;101(22):1800–1806. doi: 10.1136/heartjnl-2015-307634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Rho R.W., Patton K.K., Poole J.E., Cleland J.G., Shadman R., Anand I., Maggioni A.P., Carson P.E., Swedberg K., Levy W.C. Important differences in mode of death between men and women with heart failure who would qualify for a primary prevention implantable cardioverter-defibrillator. Circulation. 2012;126(20):2402–2407. doi: 10.1161/CIRCULATIONAHA.111.069245. [DOI] [PubMed] [Google Scholar]
- 45.Henyan N.N., White C.M., Gillespie E.L., Smith K., Coleman C.I., Kluger J. The impact of gender on survival amongst patients with implantable cardioverter defibrillators for primary prevention against sudden cardiac death. J. Intern. Med. 2006;260(5):467–473. doi: 10.1111/j.1365-2796.2006.01713.x. [DOI] [PubMed] [Google Scholar]
- 46.Ghanbari H., Dalloul G., Hasan R., Daccarett M., Saba S., David S., Machado C. Effectiveness of implantable cardioverter-defibrillators for the primary prevention of sudden cardiac death in women with advanced heart failure: a meta-analysis of randomized controlled trials. Arch. Intern. Med. 2009;169(16):1500–1506. doi: 10.1001/archinternmed.2009.255. [DOI] [PubMed] [Google Scholar]
- 47.Lee D.S., Hardy J., Yee R., Healey J.S., Birnie D., Simpson C.S., Crystal E., Mangat I., Nanthakumar K., Wang X., Krahn A.D., Dorian P., Austin P.C., Tu J.V. Clinical risk stratification for primary prevention implantable cardioverter defibrillators. Circulation: Heart Fail. 2015;8(5):927–937. doi: 10.1161/CIRCHEARTFAILURE.115.002414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Peterson P.N., Varosy P.D., Heidenreich P.A., Wang Y., Dewland T.A., Curtis J.P., Go A.S., Greenlee R.T., Magid D.J., Normand S.-L.T., Masoudi F.A. Association of single- vs dual-chamber ICDs with mortality, readmissions, and complications among patients receiving an ICD for primary prevention. JAMA. 2013;309(19):2025–2034. doi: 10.1001/jama.2013.4982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.McAlister F., Walraven C.v. External validation of the Hospital Frailty Risk Score and comparison with the Hospital-patient One-year Mortality Risk Score to predict outcomes in elderly hospitalised patients: a retrospective cohort study. BMJ Qual. Saf. 2019;28(4):284–288. doi: 10.1136/bmjqs-2018-008661. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data associated with this study have not been deposited into a publicly available repository, as the data are confidential.