Author + information
- Received January 14, 2004
- Revision received February 27, 2004
- Accepted March 2, 2004
- Published online June 16, 2004.
- Rachel Lampert, MD, FACC*,* (, )
- Craig A. McPherson, MD, FACC*,
- Jude F. Clancy, MD*,
- Teresa L. Caulin-Glaser, MD, FACC†,
- Lynda E. Rosenfeld, MD, FACC* and
- William P. Batsford, MD*
- ↵*Reprint requests and correspondence:
Dr. Rachel Lampert, Yale University School of Medicine, Section of Cardiology, 333 Cedar Street, FMP 3, New Haven, Connecticut 06520, USA.
Objectives We sought to determine whether men and women with coronary artery disease (CAD) and implantable cardioverter-defibrillators (ICDs) differ in frequency of arrhythmia recurrence and whether gender differences are independent of clinical, electrocardiographic, and electrophysiologic characteristics.
Background Epidemiologic studies show that women have a lower rate of sudden cardiac death (SCD) than men, even among patients with CAD. Whether this is due to differing susceptibilities to ischemia or to arrhythmia is unknown.
Methods The clinical records and ICD data disks of 340 men and 59 women with CAD who received an ICD between June 1990 and June 2000 were reviewed. Ventricular tachycardia (VT) or ventricular fibrillation (VF) recurrences were compared between genders and relationship with other factors was analyzed.
Results Sustained VT/VF occurred in 52% of men and 34% of women (p < 0.01). Men experienced more total VT/VF events (p < 0.01), more shock-treated VT/VF events (p < 0.03), more electrical storms (p < 0.001), and had VT/VF on more days in follow-up (p < 0.01). Gender differences were independent of measured clinical, electrocardiographic, and electrophysiologic factors. In stratified analyses, the gender differences in VT/VF recurrence were greatest in patients presenting with sustained monomorphic VT and those with inducible VT at electrophysiology study.
Conclusions Women were less likely to experience VT/VF, and had fewer VT/VF episodes, than men. These findings were strongest in patients with evidence of a stable anatomic VT circuit: those with clinical or electrophysiologically induced VT. This study suggests that differing susceptibility to arrhythmia triggering may underlie the known differences in SCD rates between men and women.
Numerous studies (1–3)of people with cardiac arrest have shown women to have a lower rate of sudden cardiac death (SCD) than men. The prevalence of coronary artery disease (CAD), the primary cause of SCD, is lower in women (4). However, even among individuals with known CAD, women are less likely to experience SCD than are men (1,2). Whether this difference is due to lower susceptibility in women to recurrent ischemia or to arrhythmia is unknown. There are few data comparing the incidence of sustained arrhythmia between men and women, and existing studies have been limited to survivors of cardiac arrest (5–8). Gender differences have been noted in inducibility of ventricular arrhythmias among patients with CAD (9,10)and in frequency of nonsustained ventricular tachycardia (VT) in patients with congestive heart failure (11), suggesting that men may have greater propensity to ventricular arrhythmia.
Patients with implantable cardioverter-defibrillators (ICDs) provide a unique opportunity to evaluate whether ventricular arrhythmia recurrence differs between men and women. In this retrospective study of patients with CAD and ICDs, we investigated whether the incidence and frequency of ventricular arrhythmia recurrence differed between men and women, and whether any such differences were due to differences in measurable clinical, electrocardiographic, or electrophysiologic factors.
Patient population and ICD devices
The records of 650 consecutive patients receiving an ICD with stored diagnostic information at Yale University-Yale New Haven Hospital between June 1990 and June 2000 were reviewed. Follow-up was complete through September 2000. Data on survival and arrhythmia recurrence were censored at the time of last follow-up. We excluded 52 patients not yet seen for routine first follow-up at two months, as presence of asymptomatic (pacer-terminated) VT could not be determined. We included only patients with Guidant (St. Paul, Minnesota) devices, model 1705 or beyond, because during the study years, these devices were the only ones that: 1) stored multiple episodes (≥128 vs. <10 for other manufacturers/models); 2) stored data on disks, allowing later retrieval of episodes; and 3) did not erase episodes on interrogation. As the large majority of devices at our institution during these years were Guidant, this resulted in the exclusion of 64 patients (<10%). Exclusion rates of men and women were similar for each exclusion criterion. Of the remaining 534 patients, 399 had CAD, based on history of documented myocardial infarction and/or results of coronary angiography, and these constitute the study population.
Clinical data were retrieved through a review of patient records. Data on ventricular arrhythmia recurrence were obtained by review of event details and electrograms stored on an ICD disk. Only episodes of VT or ventricular fibrillation (VF) requiring ICD therapy (antitachycardia pacing or shock) for termination were included in the analysis and each episode was counted once, regardless of the number of therapies received. Episodes without stored data available were not included in the analysis.
“Ventricular tachycardia/ventricular fibrillation events” were classified based on device definitions: a new arrhythmia is one occurring >30 s after a shock-terminated event or >10 s after a pacer-terminated event. “Shock-treated events” were defined as VT or VF requiring ≥1 shock for termination. “Electrical storm” was defined as the occurrence of ≥3 separate episodes of VT/VF within a 24-h period, each separated by ≥5 min, regardless of mode of termination (as defined in the Anti-arrhythmics Versus Implantable Defibrillators [AVID] trial) (12). Because repetitive VT/VF occurring on a single day could potentially skew the results, “VT/VF days,” defined as the number of calendar days on which any VT/VF occurred, were also analyzed.
Clinical and electrocardiographic variables
Clinical and electrocardiographic variables measured are listed in Table 1. Indications for ICD were categorized as either clinical sustained monomorphic VT; VF (including polymorphic VT); syncope with inducible VT at electrophysiology study (EPS); or “MUSTT/MADIT,” defined as CAD, left ventricular ejection fraction (EF) ≤40%, nonsustained VT, and inducible VT at EPS, as described in the Multicenter Unsustained Tachycardia Trial (MUSTT) (13)and the Multicenter Automatic Defibrillator Implantation Trial (MADIT) (14). Prior revascularization was categorized as “immediate” if <1 month preceding ICD implantation, or “remote.” Wall motion was determined (within one year of ICD implantation) echocardiographically in 224 patients and by left ventriculogram in 87 patients. Coronary angiography was performed in 218 patients and perfusion imaging in 193 patients. Left ventricular hypertrophy was determined by standard electrocardiogram (ECG) criteria and by Cornell voltage (68% test accuracy compared with autopsy) (15).
An EPS was performed before ICD implant, in the absence of antiarrhythmic medication, in 334 patients. Up to three extrastimuli were delivered at twice the diastolic threshold at two paced cycle lengths at two right ventricular sites (apex and outflow tract). Each extrastimulus was delivered late in systole and decremented until refractoriness or 180 ms. With the addition of subsequent extrastimuli, the preceding stimulus was set at the refractory period plus 20 ms. The protocol used was similar to that in MADIT (14,16). Studies were terminated with induction of sustained VT or VF or completion of the protocol.
Binary variables were compared between men and women by chi-square analysis. The frequency of VT/VF end points was compared using the Wilcoxon signed-rank test, as the distributions were highly skewed (Shapiro-Wilk W test p < 0.0000 for each) and using chi-square for linear trend (ordinal stages: 0, 1, 2 to 10, 11 to 100, >100 episodes). To determine the independence of associations of clinical factors with VT/VF occurrence and frequency, logistic and linear regression analyses, respectively, were performed. Associations are described by odds ratios and regression coefficients (17), respectively. Clinical factors having univariate associations with VT/VF occurrence and frequency were included as covariates for multivariable analysis. Separate models with ejection fraction as a binary and continuous variable were evaluated. A two-tailed p value of <0.05 was considered significant. Statistical analysis was performed using JMP 5.0 software (SAS Institute, Inc., Cary, North Carolina).
Clinical, electrocardiographic, and electrophysiologic data of the 340 men and 59 women are shown in Table 1. Women were older and had more diabetes and hypertension. Indications for ICD and revascularization rates were similar between genders. Few women were receiving hormone replacement therapy. Men and women underwent evaluation of wall motion, coronary anatomy, and perfusion at similar rates. Although global ventricular function was worse in men, aneurysms were more prevalent in women. Number and extent of prior infarcts, based on number of involved territories on perfusion and wall motion imaging, were similar between men and women, as was burden of CAD seen on catheterization. Although coronary anatomy, wall motion, and ECG suggested more anterior disease in women, this difference was not significant. On ECG, there were no differences between men and women in presence of prolonged QRS duration, prolonged QT interval, Q waves, or left ventricular hypertrophy.
Men and women underwent baseline EPS at similar rates, 83% and 86%. Sustained VT or VF was induced at EPS in 70% of men and 63% of women (p = 0.31). Among those with ventricular arrhythmia induced, type of arrhythmia, VT cycle length, mode of induction (number and coupling interval of extrastimuli), and response to overdrive pacing did not differ between the genders.
During average follow-up of 30 ± 22 months, sustained VT/VF requiring ICD therapy occurred in 52% of men and 34% of women (p < 0.01) (Fig. 1). Men experienced more total VT/VF events (p < 0.01) (Fig. 2). Ventricular tachycardia/ventricular fibrillation occurred on more days in men (p < 0.01), and men had more shock-treated events (p < 0.02) and more electrical storms (p < 0.001) (Table 2). After normalization of event frequencies to months of follow-up, all end points remained significantly more frequent in men (p < 0.01 for each.)
In 17 patients, all men, the number of treated episodes defined as VT or VF by the device exceeded the storage capacity of the ICD. Numbers of total confirmed VT/VF events in these patients ranged from 90 to 1,105; numbers of unconfirmed events (not included in the analyses) ranged from 297 to 5,973.
Survival after ICD implantation did not differ between men and women (Fig. 3).
Clinical, electrocardiographic, and electrophysiologic factors
In univariate analysis (Table 3), male gender, presentation with sustained VT, EF <40%, presence of aneurysm, absence of remote and immediate revascularization, use of antiarrhythmic medications at discharge, absence of beta-blockers at discharge, prolonged QRS, prolonged QT, and induction of ventricular arrhythmia at EPS were associated with greater occurrence of both any VT/VF event and frequency of VT/VF events. Diabetes was associated with a lower likelihood of recurrent arrhythmia. Other measured variables (Table 1) did not show significant univariate associations with end points. In a multivariable analysis including these factors, men remained more than twice as likely as women to undergo any VT/VF event (adjusted odds ratio 2.20, 95% confidence interval 1.02 to 4.90, p < 0.05) and had greater frequency of events (p < 0.05). In addition to gender, factors independently associated with both end points in multivariable analysis included presentation with sustained VT, inducible VT at EPS, and absence of immediate and remote revascularization.
To evaluate whether the gender difference in VT/VF was predominant in any clinical subgroups, stratified analysis was performed, initially based on indication for ICD. Men had greater occurrence and frequency of VT/VF end points among patients with a history of sustained VT and among those with “MUSTT/MADIT” criteria (13,14), but not among patients presenting with VF (Fig. 1). Also, after stratification by presence of induced arrhythmia, male gender predicted VT/VF occurrence and frequency for all end points in those patients with inducible arrhythmia, but not those without (all p < 0.05).
In this study of patients with CAD and ventricular arrhythmias treated with ICDs, gender was an important determinant of arrhythmia recurrence. Men were more likely to experience ventricular arrhythmias in follow-up and had more frequent total VT/VF events and more follow-up days on which VT/VF occurred. More severe arrhythmias (shock-treated events and electrical storms) were also more frequent in men. Male gender remained a significant predictor of VT/VF after adjustment for measured clinical, electrocardiographic, and electrophysiologic factors associated with arrhythmia recurrence. These findings were strongest in those patients with evidence of a stable anatomic VT circuit: those with clinical or induced sustained monomorphic VT.
In the Framingham study, the age-adjusted SCD risk for women was one-half that in men, and among patients with CAD one-fourth that in men, a difference not explained by risk-factor burden (1). The gender difference in SCD incidence could reflect greater male susceptibility to ischemia and/or arrhythmia. Although ischemic disease is known to occur less frequently in women (4), rates of arrhythmia recurrence have been less well described. The current findings suggest that the lower SCD rate in women with CAD may be due at least in part to a gender difference in propensity to triggering of arrhythmia.
Comparison with previous studies
Several previous studies have shown no difference in arrhythmia occurrence after ICD implantation between genders (5,7,18); one showed an increased incidence of shock in women (6). However, all studies analyzed populations of arrest or VF survivors with diverse underlying heart diseases. Similarly, in the current study, among patients presenting with VF there was no gender difference in arrhythmia recurrence.
The gender differences seen in this study cannot be explained by clinical factors shown to predict arrhythmia recurrence in patients with CAD. Consistent with previous studies (8,19), we found few clinical differences between men and women with CAD and ventricular arrhythmias. Multiple factors associated with arrhythmia occurrence or mortality in arrhythmia patients, including indication for ICD (5,7), preceding revascularization (20,21), extent (22)and site (23)of infarcted myocardium, prolonged QRS duration (24), and left ventricular hypertrophy (2,25), did not differ between genders. Although two factors potentially protective from arrhythmia were greater in women, specifically EF (a predictor of ICD shock ) and diabetes (associated with lower rates of arrhythmia in this and previous studies ), gender remained an independent predictor after adjustment for these factors. Aneurysm, linked with sustained VT (27)and arrhythmic (28)and total (29)mortality, was more prevalent in women, underscoring the fact that although arrhythmogenic substrate was similar or greater, women's arrhythmias were triggered less often. The physiologic mechanisms that underlie this phenomenon remain unknown. Possible explanations include differences in electrophysiologic properties and/or in autonomic response to triggering factors.
Cellular, animal, and human investigations have demonstrated electrophysiologic differences between genders, some dependent on circulating sex hormones, others independent of ambient hormonal milieu. Prolonged repolarization in women has been well described (30). Experimentally, estrogen affects potassium-channel protein synthesis, which alters repolarization (31)and can increase atrial refractory periods (32). However, studies of hormone replacement therapy in postmenopausal women (33)and of cyclical changes in premenopausal women (34)have not shown an influence of circulating estrogens on repolarization. Testosterone alters potassium currents in patch-clamp experiments (35)and alters the QT interval in castrated men (36). Further, electrophysiologic substrate may differ between genders even in the absence of circulating hormones. For example, myocytes isolated from female rabbits show reduced outward potassium current density (37)and longer action potential duration (38). However, in the current study no differences were seen in ECG measures of repolarization (QT/QTc). It is likely that the extent of myocardial disease became a more important determinant of QT/QTc than gender.
Previous data are conflicting with regard to inducibility of VT during EPS in women with CAD (8–10,39). In this population, with history of clinical or induced arrhythmia, prevalence of inducible VT was not significantly greater in men. Although inducible arrhythmia was a strong determinant of future VT/VF events in this study and others (7,40), adjustment for inducibility did not alter the associations of gender and arrhythmia. Surprisingly, it was those patients with the most homogeneous profile, clinical sustained monomorphic VT and/or inducibility of VT at EPS, in whom gender most strikingly predicted spontaneous VT/VF recurrence. This suggests that it is differences in susceptibility to arrhythmia triggering, not in the underlying electrophysiologic substrate, that mediate the gender differences in arrhythmia recurrence.
Previous studies have demonstrated gender differences in autonomic nervous system function. Resting vagal tone, as measured by heart rate variability, is higher in women in most (41,42), although not all (43), studies. This may be due to circulating hormones (41), as estrogen alters norepinephrine release (44). However, the gender difference in vagal tone persists in elderly women (42), suggesting that the autonomic differences between men and women may be intrinsic. Also, women have less sympathetic response to induced mental stress (45). As sympathetic activation can influence arrhythmia occurrence and severity (46,47), differences in autonomic response may underlie the gender differences in arrhythmia seen in the current study.
Although women had fewer arrhythmia recurrences than men, a substantial number did have potentially life-threatening ventricular arrhythmias treated by the ICD, and survival after ICD implantation did not differ between genders. Clearly, both genders benefit from ICD therapy, as found in the AVID trial (48).
The ICDs in this study had the largest storage capabilities available, but in some patients tachyarrhythmia episodes exceeded device memory, resulting in the potential loss of data. As these episodes occurred only in men, the current findings are likely an underestimation of the magnitude of the gender difference. Also, in this retrospective study, not all patients had undergone all diagnostic assessments. However, men and women underwent EPS and imaging studies in similar proportions, making systematic bias unlikely. Functional capacity was not tested in most patients, and New York Heart Association functional class, correlated in univariate analysis with appropriate ICD shock in previous studies (although multivariate analysis was not performed in these studies [49,50]), was not recorded consistently. However, because structural heart disease as measured by multiple imaging studies was similar in men and women, it is unlikely that potential unmeasured differences in New York Heart Association functional class are the primary source of the gender difference in arrhythmia. Future investigation of congestive heart failure as a potential trigger for arrhythmia in men and women is warranted. Also, postmortem ICD interrogation was not performed. Whether gender differences exist in arrhythmia recurrence near the patient's end of life requires further investigation. As in previous studies (8,13,14,39,48), there were relatively small numbers of women in our population, but the gender difference remained significant in evaluation of multiple end points and controlling for multiple potential confounders.
This study demonstrates a significant difference in arrhythmia occurrence and frequency between men and women with CAD and ICDs. Women were less likely than men to experience ventricular arrhythmias in follow-up and had fewer arrhythmias when they did recur. These gender differences were not explained by measured clinical, electrocardiographic, or electrophysiologic factors. Determining the mechanisms underlying these differences requires further investigation. This study suggests that differing susceptibility to arrhythmia triggering may underlie the known differences in SCD rates between men and women.
- Anti-arrhythmics Versus Implantable Defibrillators trial
- coronary artery disease
- ejection fraction
- electrophysiologic study
- implantable cardioverter-defibrillator
- Multicenter Automatic Defibrillator Implantation Trial
- Multicenter UnSustained Tachycardia Trial
- sudden cardiac death
- ventricular fibrillation
- ventricular tachycardia
- Received January 14, 2004.
- Revision received February 27, 2004.
- Accepted March 2, 2004.
- American College of Cardiology Foundation
- Grimm W.,
- Flores B.,
- Marchlinski F.
- Albert C.M.,
- McGovern B.A.,
- Newell J.B.,
- Ruskin J.N.
- Buxton A.,
- Hafley G.,
- Lehmann M.,
- et al.
- Exner D.,
- Pinski S.,
- Wyse G.,
- et al.
- Casale P.,
- Devereux R.,
- Alonso D.,
- Campo E.,
- Kligfield P.
- ↵Feinstein A. Principles of Medical Statistics. Boca Raton, FL: Chapman and Hall/CRC, 2002:370–3.
- Ruppel R.,
- Schluter C.A.,
- Boczor S.,
- et al.
- Every N.,
- Fahrenbruch C.,
- Hallstrom A.,
- Weaver D.,
- Cobb L.
- Borger van der Burg A.,
- Bax J.,
- Boersma E.,
- Pauwels E.,
- van der Wall E.,
- Schalij M.
- Stone P.,
- Raabe D.,
- Jaffe A.,
- et al.
- Fadl Y.,
- Zareba W.,
- Moss A.,
- MDPIT Investigators
- Simpson R.,
- Cascio W.,
- Crow R.,
- Schreiner P.,
- Rautaharju P.,
- Heiss G.
- Drici M.,
- Burklow T.,
- Haridasse V.,
- Glazer R.,
- Woosley R.
- ↵Liu XK, Katchman A, Drici MD, et al. Gender difference in the cycle length dependent QT and potassium currents in rabbits. J Pharmacol Exp Ther 1998;285:672–9
- Huikuri H.,
- Pikkujamsa S.,
- Airaksinen J.,
- et al.
- Ryan S.,
- Goldberger A.,
- Pincus S.,
- Mietus J.,
- Lipsitz L.
- Umetani K.,
- Singer D.H.,
- McCraty R.,
- Atkinson M.
- Du X.,
- Riemersma R.,
- Dart A.
- Lampert R.,
- Jain D.,
- Burg M.,
- Batsford W.,
- McPherson C.
- Shusterman V.,
- Aysin B.,
- Gottipaty V.
- Pacifico A.,
- Ferlic L.,
- Cedillo-Salazar R.,
- Nasir N.,
- Doyle T.,
- Henry P.
- Trappe H.,
- Wenzlaff P.,
- Pfitzner P.,
- Fieguth H.