Sex Differences in the Risk of First and Recurrent Ventricular Tachyarrhythmias Among Patients Receiving an Implantable Cardioverter-Defibrillator for Primary Prevention.

Importance: Current guidelines for primary implantable cardioverter-defibrillator (ICD) therapy do not account for sex differences in arrhythmic risk in ICD candidates. Objective: To evaluate the association between sex and risk of ventricular tachyarrhythmia (VTA) and mortality. Design, Setting, and Participants: This cohort study compared differences in the risk of VTA and mortality between 4506 men and women enrolled in the 4 Multicenter Automatic Defibrillator Implantation Trials (MADIT) between July 1, 1997, and December 31, 2011. Data from prospective randomized controlled multicenter studies were analyzed retrospectively. Men and women with an ICD or cardiac resynchronization therapy defibrillator who were enrolled in all MADIT studies were included. Data were analyzed between January 10 and June 10, 2021. Exposures: ICD implant. Main Outcomes and Measures: The primary end point was sustained VTA, defined as ICD-recorded, treated or monitored VTA at least 170/min or ventricular fibrillation. Secondary VTA end points included VTA at least 200/min, appropriate ICD shocks, and appropriate antitachycardia pacing. All end points were included in a first and recurrent event analysis.
Results: Of the 4506 study participants, 3431 were men (76%). Mean (SD) age of the cohort was 64 (12) years. For women vs men, the mean (SD) age (64 [12] years vs 64 [11] years) and left ventricular ejection fraction (24% vs 25%) were similar, but women exhibited a higher frequency of nonischemic cardiomyopathy (454 of 1075 women [42%] vs 2535 of 3431 men [74%]). Women had significantly lower 3-year cumulative probability of sustained VTA (16% vs 26%), fast VTA (9% vs 17%), and appropriate ICD shocks (7% vs 15%) compared with men (P < .001 for all). Multivariable analysis showed that female sex was independently associated with at least 40% lower risk of all first and recurrent VTA end points (P < .001 for all), including the primary end point (first event, HR = 60 [95% CI, 50-73], P < .001; recurrent event, HR = 49 [95% CI, 43-55], P < .001), after accounting for the competing risk of all-cause mortality and nonarrhythmic mortality. The lower VTA risk associated with female sex was consistent in risk subsets but was significantly more pronounced in patients with nonischemic cardiomyopathy (female vs male in the ischemic group: hazard ratio, 0.73 [95% CI, 0.56-0.95], P = .02; nonischemic group: hazard ratio, 0.50 [95% CI, 0.38-0.66], P < .001; P = .03 for interaction between female sex and cardiomyopathy). Conclusions and Relevance: Findings suggest that women display a significantly lower risk of first and recurrent life-threatening VTA events than men, and that it is more pronounced in patients with nonischemic cardiomyopathy, suggesting a need for sex-specific risk assessment for primary prevention ICD therapy.

Introduction

Differences due to sex have been known to have strong correlations with development and progression of disease in patients with heart failure (HF).[1,2] Prior studies of patients with an implantable cardioverter-defibrillator (ICD) have demonstrated conflicting results regarding the association of sex with the incidence of appropriate device therapy, overall mortality, and thus the benefit of prophylactic placement of ICDs in women with HF. Several studies have reported a lower incidence of ICD therapy among female patients,[3,4,5,6] whereas others have reported no difference in risk of appropriate ICD therapy between sexes.[7,8,9] Regarding overall mortality and ICD benefit, some studies have found a lower risk of mortality among women and have suggested that women derive a smaller ICD benefit compared with men.[3,7,9] However, other reports showed a similar mortality risk among men and women with prophylactic ICDs.[8,10,11]

These contradictory findings may be attributed to the fact that women are underrepresented in all major ICD randomized clinical trials, generally accounting for less than 30% of enrolled patients.[1,12,13,14,15] Likewise, implantation of an ICD has been shown to be significantly less common in women than in men.[13,14,16] Nonischemic cardiomyopathy (NICM) is reported more frequently than ICM in women than in men as the underlying etiology of cardiomyopathy, which may affect the arrhythmic risk when male and female patients are compared.[17]

In this study, we aimed to evaluate the association between sex and the risk of first occurrence of sustained ventricular tachyarrhythmia (VTA), total VTA and shock burden during the follow-up period in a recurrent event analysis, and nonarrhythmic mortality among all patients with primary prevention ICD implantation who were enrolled in the landmark Multicenter Automatic Defibrillator Implantation Trials (MADIT). Outcomes by sex were further stratified by the type of cardiomyopathy.

Methods

Study Population

This cohort study included 4506 patients with an ICD or cardiac resynchronization therapy defibrillator (CRTD) who were enrolled in the 4 MADIT studies between July 1, 1997, and December 31, 2011. Design, study protocol, and results of each of these trials have been published previously. MADIT II enrolled 730 patients with post–myocardial infarction who had a left ventricular ejection fraction (LVEF) of 30% or less and were randomized to ICD therapy vs conventional therapy.[18] MADIT-RISK was a prospective observational study that enrolled 466 patients similar to those in MADIT II.[19] MADIT-CRT enrolled 1814 patients with ICM or NICM, an LVEF of 30% or less, a QRS duration greater than or equal to 130 ms, and New York Heart Association functional class I or II. Patients were randomly assigned in a 3:2 ratio to have CRTD implantation (n = 1083) or ICD implantation alone (n = 731).[20] MADIT-RIT enrolled 1465 patients, all of whom met guideline criteria to have ICD or CRTD implantation for primary prevention.[21] The rate of missing data from these studies was less than 5%.

Each study was approved by the institutional review boards at each enrolling site before participation. All patients provided written informed consent before enrollment into the respective trials. The current analysis of this cohort study was performed in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Definitions and End Points

In all MADIT trials, device interrogations were reviewed and adjudicated blindly by an independent committee of at least 2 experienced electrophysiologists and cardiologists using identical definitions. Details of device programming and VTA definitions are provided in eTable 1 in the Supplement. In all MADIT studies, programming was similar for men and women.

The primary end point of the present study was the first occurrence of VTA, defined as ICD-recorded, treated or monitored, sustained ventricular tachycardia at least 170/min or ventricular fibrillation. Secondary VTA end points included fast VTA at least 200/min, appropriate ICD shocks, and appropriate ICD antitachycardia pacing. Ventricular tachyarrhythmia was defined by the adjudication committee according to combined evaluation of frequency, QRS morphology, and regularity of the rhythm. A ventricular tachycardia event needed to last for more than 30 seconds or, if therapy was administered, until termination of the arrhythmia (including polymorphic ventricular tachycardia). Detection zone was 145/min in all studies. To account for the competing risk of death, we evaluated the following mortality outcomes: all-cause mortality and nonarrhythmic mortality (defined as death without experiencing sustained VTA at any time during follow-up after ICD implantation).

Statistical Analysis

Categorical data are summarized as frequencies and percentages. Continuous data are summarized as mean (SD) values. Baseline characteristics were compared between the male and female subgroups with the Wilcoxon rank sum test (for continuous variables) and the χ2 test (for categorical variables).

Kaplan-Meier cumulative probabilities were used to graphically display the association of female and male sex with the end point of VTA over time. The log-rank test was used for determination of statistical significance.

Multivariable Cox proportional hazards regression analysis was used to identify and evaluate the association between sex and the risk of VTA and the other relevant end points. For the multivariable adjustment, relevant clinical covariates were determined according to the results of a stepwise selection process and were included if statistically significant at P < .05 in multivariable models for either end point (VTA or death). Potential risk factor candidates are shown in eTable 2 in the Supplement.

Model 1 used a dichotomized functional form of continuous variables to support the goal of developing a relatively simple and easily implemented scoring method. Thresholds for categorization of numeric variables were prespecified with clinically well-accepted criteria. Models were adjusted for study type by stratification. In model 2, the multivariable model derivation process is identical except only continuous variables were used rather than dichotomized forms.

To validate the results and to avoid bias associated with covariates, a propensity score model for predicting female using logistic regression for variables that met the P < .05 significance criteria in the multivariate model was created. The following independent predictors were identified: African American race, smoking, diastolic blood pressure, New York Heart Association class, history of HF hospitalization, history of atrial tachyarrhythmia, and the presence of ICM. These predictors were included in the propensity score model. Each participant in the study was assigned a propensity score quintile value, 1 to 5, and this quintile of the score was adjusted for in the model via stratification.

The association of sex with VTA risk was evaluated in prespecified subgroups of interest, using interaction-term analysis. Within each group (women and men), we used cumulative incidence function curves to estimate the probability of first VTA as the event of interest, and the probability of death without prior VTA as a competing risk.

The Anderson-Gill non–gap time model was used to estimate the hazard ratios (HRs) for recurrent VTA events, recurrent fast VTA at least 200/min, recurrent appropriate ICD shocks, and recurrent appropriate ICD antitachycardia pacing. Models were adjusted for covariates as mentioned earlier. Mean cumulative event rate was presented with Ghosh-Lin curves to display mean number of recurrent events per patient as appropriate.

All tests of significance were 2-tailed, with P < .05 indicating statistical significance. Data analysis was carried out with SAS, version 9.4 (SAS Institute). Data were analyzed between January 10 and June 10, 2021.

Results

Baseline Characteristics

The study included 4506 patients, 1075 women (24%) and 3431 men (76%). Mean (SD) age of the cohort was 64 (12) years. A total of 517 patients were Black or African American (11%), 207 were Hispanic (5%), and 3796 were White (84%). Race and ethnicity were self-reported by the participants. Baseline characteristics of the study population by sex are shown in Table 1. For women vs men, left ventricular ejection fraction (24% vs 25%) was similar, but women exhibited a lower frequency of ischemic cardiomyopathy (454 of 1075 women [42%] vs 2535 of 3431 men [74%]). Compared with men, women were more likely to be Black or African American, to have more advanced HF symptoms, to have left bundle-branch block, and to be treated with CRTD. However, they were less likely to have ICM, have a past myocardial infarction, have a history of smoking, or be treated with amiodarone. Men and women had similar age, LVEF, body mass index, and vital signs at enrollment (Table 1).

Table 1.

Baseline Clinical Characteristics

Clinical characteristic	Men (n = 3431)	Women (n = 1075)	P value
Age, mean (SD), y	64 (11)	64 (12)	.30
Race and ethnicity, No. (%)a
Black or African American	310 (9)	207 (19)	<.001
Hispanic	155 (5)	52 (5)	.600
White	2973 (87)	823 (77)	<.001
BMI, mean (SD)	29 (6)	29 (7)	.19
BP, mean (SD), mm Hg
Systolic	123 (18)	123 (19)	.33
Diastolic	72 (11)	71 (11)	<.001
Heart rate, mean (SD), /minb	70 (12)	72 (12)	<.001
% LVEF, mean (SD)	25 (6)	24 (6)	.08
CRTD, No. (%)	1257 (37)	516 (48)	<.001
LBBB, No. (%)b	1225 (51)	570 (75)	<.001
QRS duration, mean (SD), msb	146 (31)	149 (26)	.009
NYHA II-IV, No. (%)	2815 (82)	1000 (93)	<.001
Ischemic cardiomyopathy, No. (%)	2535 (74)	454 (42)	<.001
Hypertension, No. (%)	2180 (64)	724 (67)	.02
History of atrial tachyarrhythmia, mean (SD)	115 (11)	38 (9)	.22
History of NSVT, No. (%)	240 (7)	54 (5)	.02
Diabetes, No. (%)	1083 (32)	355 (33)	.32
Current cigarette use, No. (%)	523 (15)	126 (12)	.006
MI before enrollment, No. (%)	2190 (64)	381 (35)	<.001
Creatinine, mean (SD), mg/dLb	1.2 (0.4)	1.0 (0.4)	<.001
Aldosterone, No. (%)b	862 (25)	328 (31)	.049
Amiodarone, No. (%)	246 (7)	36 (3)	<.001
ACE inhibitor or ARB, No. (%)	3112 (91)	971 (90)	.71
Aspirin, No. (%)	2387 (70)	656 (61)	<.001
β-Blocker (excluding sotalol), No. (%)	3000 (87)	978 (91)	<.001
Digitalis, No. (%)	849 (25)	306 (28)	.01

Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin 2 receptor blocker; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); BP, blood pressure; CRTD, cardiac resynchronization therapy defibrillator; LVEF, left ventricular ejection fraction; LBBB, left bundle-branch block; MI, myocardial infarction; NYHA, New York Heart Association functional class; NSVT, nonsustained ventricular tachycardia.

Fourteen patients reported race and ethnicity as both Hispanic and White.

Calculated with available data only.

Sex and the Risk of First VTA

During a mean follow-up of 3 years, a total of 853 of 4506 patients (19%) had at least 1 episode of VTA. The 3-year probability of VTA was significantly lower among women than among men (VTA, 16% vs 26%; fast VTA, 9% vs 17%; and appropriate ICD shocks, 7% vs 15%, respectively; P < .001 for each comparison for the overall difference at any point during follow-up) (eFigure 1 in the Supplement).

Consistent with the univariate findings, multivariable Cox modeling (Table 2) also showed a significantly lower risk of first VTA event associated with female sex: 40% lower risk of sustained VTA, 45% lower risk of fast VTA at least 200/min, 44% lower risk of appropriate ICD shock, and 39% lower risk of antitachycardia pacing (P < .001 for all). Similar results were observed for model 2 (eTable 3 in the Supplement).

Table 2.

Multivariable Cox Regression (for First Event) and Anderson-Gill Model (for Recurrent Events) Evaluating the Association of Female vs Male Sex on the Development of First and Recurrent Ventricular Tachyarrhythmia (VTA) End Points

End point	All patients		Ischemica		Nonischemica
	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
First VTA event
VTA	0.60 (0.50-0.73)	<.001	0.73 (0.56-0.95)	.02	0.50 (0.38-0.66)	<.001
VTA ≥200/min	0.55 (0.42-0.70)	<.001	0.69 (0.49-0.98)	.04	0.42 (0.30-0.60)	<.001
Shock	0.56 (0.42-0.74)	<.001	0.68 (0.47-0.97)	.03	0.44 (0.28-0.69)	<.001
ATP	0.61 (0.49-0.77)	<.001	0.81 (0.58-1.12)	.19	0.49 (0.36-0.67)	<.001
Recurrent VTA events
VTA	0.49 (0.43-0.55)	<.001	0.56 (0.47-0.67)	<.001	0.44 (0.37-0.51)	<.001
VTA ≥200/min	0.50 (0.41-0.60)	<.001	0.64 (0.49-0.83)	<.001	0.39 (0.30-0.52)	<.001
Shock	0.52 (0.42-0.64)	<.001	0.60 (0.47-0.78)	<.001	0.43 (0.30-0.60)	<.001
ATP	0.46 (0.40-0.53)	<.001	0.55 (0.44-0.69)	<.001	0.40 (0.33-0.48)	<.001

Abbreviations: ATP, antitachycardia pacing; HR, hazard ratio.

Interaction models using the entire population with an interaction between the cardiomyopathy and female sex. Model 1 was adjusted for race, device type, current smoking, systolic blood pressure, ejection fraction, past percutaneous coronary intervention, past myocardial infarction, and New York Heart Association functional class.

Propensity score models stratified for baseline independent variables (Methods section) demonstrated consistent results (HR, 0.46 [95% CI, 0.41-0.53]; P < .001 for VTA) (eTable 4 in the Supplement). Subgroup analysis by age, LVEF, presence of diabetes, ICM vs NICM, and device type (ICD alone vs CRTD) demonstrated that the lower risk of sustained VTA events associated with female sex was consistent in all subgroups (Figure 1). Although CRTDs and ICDs have different arrhythmic benefits and risks, the risk of first VTA was equivalent between the 2 subgroups. Also, the interaction between sex and HF etiology was statistically significant, wherein among patients with NICM female sex was associated with a pronounced 52% (P < .001) lower risk of a first VTA, whereas among patients with ICM there was a corresponding 27% (P = .02) lower VTA risk (P = .03 for sex–heart failure etiology interaction).

Figure 1.

Association of Sex With Risk of Ventricular Tachyarrhythmia in Subgroups

Forest plot with multivariable Cox regression evaluating the association of female vs male sex with the development of sustained VTA, according to preselected subgroups. CRTD indicates cardiac resynchronization therapy defibrillator; EF, ejection fraction; HR, hazard ratio; ICD, implantable cardioverter-defibrillator; and VTA, ventricular tachyarrhythmia.

VTA Risk by Cardiomyopathy

Because the interaction model for the primary end point showed that the association of female sex with a lower VTA risk was significantly more pronounced among NICM patients, we further report all arrhythmic outcomes for each cardiomyopathy type separately. Results remained consistent, demonstrating that in both cardiomyopathy types the 3-year risk of any sustained VTA and fast VTA was significantly lower in women than in men (VTA in NICM was 14% in women vs 27% in men, P < .001; VTA in ICM was 19% in women vs 26% in men, P = .03). However, sex differences in arrhythmic risk were more pronounced in NICM patients (Figure 2A and B) than in ICM patients (Figure 2C and D). Furthermore, in NICM patients, appropriate shocks, antitachycardia pacing, or both were significantly lower in women than in men (shock: 5% in women vs 14% in men, P < .001; antitachycardia pacing: 12% in women vs 21% in men, P < .001), whereas in ICM, female sex was associated with a significantly lower risk for shocks (12% in women vs 16% in men; P = .04) and a trend for lower risk for antitachycardia pacing (13% in women vs 16% in men; P = .16) (eFigure 2 in the Supplement). Multivariate Cox models (model 1) showed that in ICM patients, female sex was associated with a 37% risk reduction for VTA (HR, 0.73 [95% CI, 0.56-0.95]; P = .02) whereas in NICM women had half the risk of men for VTA (HR, 0.50 [95% CI, 0.38-0.66]; P < .001) (Table 2).

Figure 2.

Cumulative Probability of Ventricular Arrhythmia by Sex and Cardiomyopathy

Three-year Kaplan-Meier cumulative probability by sex for sustained VTA and fast VTA at least 200/min in nonischemic patients (panels A and B, respectively) and ischemic patients (panels C and D, respectively). HR indicates heart rate; VTA, ventricular tachyarrhythmia.

Sex and the Risk of VTA Burden

During follow-up, a total of 2581 VTA events occurred in all patients. The mean number of all VTA events per patient was significantly lower in women compared with men. At 3 years, the mean cumulative number of sustained VTA events per patient was lower for women compared with men in NICM (0.45 events per patient in women vs 0.93 events per patient in men; P < .001) and ICM (0.45 events per patient in women vs 0.79 events per patient in men; P < .001) (Figure 3). Consistently, multivariate analysis showed that the risk of recurrent VTA was significantly lower among women compared with men with both ICM (HR, 0.56 [95% CI, 0.47-0.67]; P < .001) and NICM (HR, 0.44 [95% CI, 0.37-0.51]; P < .001) (Table 2).

Figure 3.

Risk of Ventricular Tachyarrhythmic Burden by Sex and Cardiomyopathy

Mean cumulative rate for recurrent events per patient, stratified by sex for sustained ventricular tachyarrhythmia and fast ventricular tachyarrhythmia at least 200/min in nonischemic patients (panels A and B, respectively) and ischemic patients (panels C and D, respectively).

Competing Risk of Death

All-cause mortality occurred in 403 out of 4506 patients (9%), without any significant difference between women and men (adjusted HR, 0.80 [95% CI, 0.65-1.11]; P = .23). The cumulative incidence function curves in both the male and female groups demonstrated that the risk of VTA was significantly higher than the competing risk of nonarrhythmic mortality, but the difference was less pronounced among women. In women with NICM, the 3-year rates of VTA vs nonarrhythmic mortality were 14% vs 7%, respectively (P < .001). Similarly, in women with ICM respective rates were 20% vs 11% (P < .001; eFigure 3A-D in the Supplement).

Discussion

In the present study, we aimed to assess sex differences in first and recurrent VTA events after prophylactic ICD and CRTD placement, using a combined database of all MADIT trials, which collectively enrolled 4506 patients and included 1075 women. To our knowledge, this is the largest cohort of women analyzed from uniformly adjudicated arrhythmia and mortality data of primary prevention ICD randomized controlled trials. Our findings show that (1) women have a significantly lower risk of first VTA occurrence, as well as a significantly lower VTA burden after prophylactic ICD placement; (2) the statistically significant differences in VTA risk between women and men were observed in all risk subsets analyzed, including age, LVEF, presence of diabetes, HF etiology, and device type; (3) the lower risk observed in women is more significantly pronounced in patients with NICM than with ICM; and (4) both women and men experience a higher risk of VTA events vs nonarrhythmic mortality after ICD implantation, indicating potential lifesaving benefit of primary ICD therapy regardless of sex.

Sex Differences in VTA Events in HF Patients

Prior studies have yielded conflicting results regarding sex differences in the incidence of appropriate ICD therapy among women. Substudies derived from randomized clinical trials of primary prevention ICDs (DEFINITE and SCD-HeFT) found no significant difference between men and women in incidence of appropriate therapies.[8,9] These findings have also been observed in a large observational registry.[7] Conversely, other substudies of major ICD trials (MUSTT and MADIT II) have previously demonstrated that the risk for appropriate ICD therapy was significantly lower in women than in men.[6,10] In addition, a lower incidence of appropriate ICD therapies among women has been noted in meta-analyses and observational studies.[3,4,5,7,11] Our findings align with those of the latter group, but unlike previous substudies, the present study has comparably more power owing to a larger number of enrolled women, and we assessed the difference in VTA events within clinically relevant risk subsets.

Sticherling et al[22] reported that in a contemporary cohort of 5033 European recipients of a primary preventive ICD, the risk for women to have an appropriate ICD shock was 39% lower compared with that for men. Our study confirmed these findings and expanded them to include adjudicated VTA events, treated and untreated. In this study we report on slow VTA and appropriate antitachycardia pacing, whereas the European study examined appropriate ICD shocks only, which are known to be associated with the programming and usually reflect fast VTAs.

Although patients with both cardiomyopathies demonstrated lower risk of VTA among women, the subgroup analysis in the present study demonstrated that this was especially pronounced in patients with NICM compared with those with ICM. Women were more likely to receive CRT and to have underlying left bundle-branch block. However, separate analysis of 1380 patients enrolled in MADIT-CRT with follow-up echocardiography at 12 months showed that the mean percentage of improvement ([LVEF at 12 months − LVEF at enrollment]/LVEF at enrollment] × 100) was 7% in men and was similar to the mean percentage of improvement observed in women (8%). This suggests that the protective effect associated with female sex is not associated with enhanced improvement in LVEF but rather is multifactorial. Attenuating estrogen effects on development of cardiac fibrosis in women along with harmful effects of androgens in this process, regardless of cardiomyopathy etiology, may be an alternative mechanism that can help explain the observed lower risk of VTA in women.[23,24] In addition, several studies have shown calcium leak from the sarcoplasmic reticulum to be lower in the cardiomyocytes of women with HF compared with men, which may result in a lower risk of developing malignant ventricular arrhythmias.[24,25]

Imaging has a possible role as a tool for risk stratification. It has been a subject of interest in studies focused on developing individualized risk stratification strategies and thus may be an important consideration in the identification of sex-specific risk factors for selection of patients for primary ICD implantation.[26]

Sex Differences in the Benefit of Primary ICD Implantation

Prior studies have yielded controversial results regarding mortality and ICD benefit for women in the population with primary prevention ICD implantation. To assess the potential benefit of primary ICD implantation in women and men, we evaluated the difference in the risk between VTA (in which prophylactic ICD implantation is potentially lifesaving) and the risk of nonarrhythmic death (in which prophylactic ICD implantation will not be helpful).[25,27] As shown in eFigure 2 in the Supplement, although the difference in the risk between VTA and nonarrhythmic mortality was attenuated in women vs men, our findings show that in women the risk for VTA remained significantly higher than that for nonarrhythmic mortality. These data stress the clinically important survival benefit of primary ICD implantation in women.

Sex Differences in VTA Burden

Understanding of the risk factors, complications, and treatments of VTA is based mostly on studies that have evaluated VTA in a binary fashion (present or absent or first recurrent) and have not investigated VTA as a burden. Nevertheless, several studies have demonstrated that increased VTA burden and frequent ICD therapy are associated with HF hospitalization, death, and significant deterioration in a patient’s quality of life and mental health.[28,29,30,31]

To our knowledge, this is the first study to examine sex differences in not only initial occurrence of VTA or first appropriate ICD therapy but also the overall burden of each of these end points among patients with primary prevention ICD implantation. More specifically, our findings depicted that women have approximately half the risk of recurrent VTA, or recurrent appropriate ICD shocks, compared with men, which was again more pronounced in NICM than in ICM.

Limitations

Our study is subject to several limitations. The current analysis is a post hoc and nonprespecified analysis; thus, it should be considered a hypothesis-generating study only. First, although our models were adjusted for significant predictors, which helps reduce uncertainty when the association between sex and the hazard of VTA or death is estimated, that it is not a randomized controlled trial may have limited our ability to reduce bias that may have been associated with differences in the frequency of these predictors between women and men. Second, the association of recently taking drugs (including sodium-glucose cotransporter 2 inhibitors and sacubitril/valsartan) with VTA risk in the HF population remains unknown, which may have affected our results. Nevertheless, the benefit of these drugs seems to be independent of sex and should not affect our conclusion significantly.[32,33] Third, our cohort included patients with HF due to reduced LVEF, and therefore the results may not be applicable for other types of patients with HF. Fourth, the eligibility criteria and the exclusion of elderly patients or those with advanced diseases (such as patients with creatinine level >2.5 mg/dL) in all our trials limits the generalizability of these findings. Yet the inclusion of very different ICD populations in the trials still provides us with a relatively mixed group of patients, allowing easier adoption of the results in the general population.

Conclusions

Our combined data from landmark ICD trials, using uniform arrhythmia and mortality adjudication data, show that in both ICM and NICM, female patients have a lower likelihood of first VTA events and overall VTA burden compared with male patients, with no difference in the risk of all-cause mortality. However, the sex difference in VTA risk was significantly higher in NICM than in ICM patients. Nevertheless, in women, the risk for VTA remains higher than the risk of nonarrhythmic mortality, suggesting that women with both ICM and NICM derive benefit from primary prevention ICD implantation.