Nonsustained Ventricular Tachycardia in Severe Heart Failure
Independent Marker of Increased Mortality due to Sudden Death
Background The goal of the study was to determine the prognostic value of nonsustained ventricular tachycardia (NSVT) in total mortality in severe congestive heart failure (CHF) and in death modes. NSVT is associated with an increased mortality in CHF. However, the predictive value of NSVT as a marker for sudden death or death due to progressive heart failure has not been determined.
Methods and Results Five hundred sixteen patients from the GESICA trial (33.4% with NSVT) were initially studied with the results of 24-hour Holter and 2 years of follow-up. Within 2 years, 87 of 173 patients (50.3%) with NSVT and 106 of 343 patients (30.9%) without NSVT died. Relative risk (RR) was 1.69 (95% confidence interval [CI], 1.27 to 2.24; P<.0002), and Cox proportional hazard analysis was 1.62 (95% CI, 1.22 to 2.16; P<.001). Sudden death increased from 8.7% (30 of 343) to 23.7% (41 of 173) in patients with NSVT (RR, 2.77; 95% CI, 1.78 to 4.44; P<.001). Progressive heart failure death was also increased from 17.5% (60 of 343) to 20.8% (36 of 173) (P=.22). Quantitative analysis of 24-hour Holter (first 295 patients) demonstrated that couplets had a similar RR to that of NSVT for both total mortality (RR, 1.81; 95% CI, 1.22 to 2.66; P<.002) and sudden death (RR, 3.37; 95% CI, 1.57 to 7.25; P<.0005). Couplets and/or NSVT (ventricular repetitive beats) were even more predictive for sudden death (RR, 10.1; 95% CI, 1.91 to 52.7; P<.01).
Conclusions In patients with CHF, NSVT is an independent marker for increased overall mortality rate and sudden death. The absence of NSVT and ventricular repetitive beats in a 24-hour Holter indicates a low probability of sudden death.
Despite recent remarkable progress in the treatment of congestive heart failure, mortality continues to be extremely high. Although it is not easy to accurately determine the mechanism causing death in patients with severe heart failure, it is generally accepted that ≈40% of these patients die suddenly.1 2 3 4 5 6 Ventricular tachycardia and fibrillation have been reported as the culprits in ≈80% of ambulatory patients in whom Holter recordings were being obtained at the time of death.7 8 Furthermore, studies demonstrating a high prevalence of ventricular arrhythmias in patients with heart failure2 5 6 9 10 11 12 have led to the clinical knowledge that the presence of ventricular arrhythmias may represent an additional marker of the severity of left ventricular dysfunction.5 12 In addition, studies have demonstrated the independent value of the presence of complex ventricular arrhythmias—couplets and nonsustained ventricular tachycardia (NSVT)—as predictors of total mortality.6 10 11 13 However, to date, the relation between the presence of ventricular rhythm disturbances and the mechanism of death in patients with left ventricular dysfunction remains controversial.
In the recently reported GESICA trial,14 the effect of low-dose amiodarone therapy on total mortality was randomly evaluated in patients with severe chronic heart failure. Our initial hypothesis that the antiarrhythmic effects of amiodarone might result in reduced incidence of sudden death led us to perform a stratified randomization protocol according to the presence or absence of NSVT on a 24-hour Holter performed at the time of patient enrollment.
The aim of this study was to determine in a homogeneous population of patients with advanced chronic heart failure the independent prognostic value of NSVT on total mortality and its relation to the death mechanisms. As such, this study presents data concerning the second main objective of the GESICA trial.14
This investigation was a prospective cohort study designed to evaluate the prognostic value of the presence or absence of NSVT in 24-hour Holter recordings obtained before randomization in 516 patients with severe heart failure and marked left ventricular systolic dysfunction.
Selection of Patients
Inclusion criteria were (1) advanced chronic heart failure adequately treated with a low-sodium diet, diuretics, digitalis, and ACE inhibitors; (2) the absence of antiarrhythmic treatment; and (3) marked left ventricular systolic dysfunction. Patients had to be stable and in an advanced functional capacity as determined on the basis of a combination of the Canadian Society criteria and the Specific Activity Scale, which was equivalent to New York Heart Association advanced classes II, III, and IV.14 Additionally, a minimum of at least two of the following three indexes of systolic myocardial dysfunction had to be present: chest radiograph cardiothoracic ratio of >0.55, radionuclide ejection fraction of ≤0.35, and an end-diastolic diameter as measured by echocardiography of ≥3.2 cm/m2.
Exclusion criteria included treatment with amiodarone during the 3 months before randomization, thyroid dysfunction, severe respiratory failure, concomitant serious associated clinical diseases, severe valvular heart disease (ie, mitral stenosis, aortic stenosis), and/or hypertrophic or restrictive cardiomyopathy; angina pectoris, myocardial infarction, onset of heart failure, or history of syncope within the past 3 months; atrioventricular conduction disorder, history of sustained ventricular tachycardia or ventricular fibrillation; and asymptomatic ventricular tachycardia for >10 beats with an RR interval of <600 msec (rate of >100 min).
Group Definition, Follow-up, and Evaluation of Results
Forms containing clinical information as well as Holter tapes and reports were submitted by individual investigators to the coordinating center for data entry and subsequent analysis. The definition of NSVT on the 24-hour admission Holter required the presence of three or more consecutive ventricular beats at a rate ≥100 bpm. The Holter tapes of all 516 patients enrolled into the trial were reviewed by a committee to confirm the presence or absence of NSVT. In addition, in the first 300 enrolled patients, the total amount of premature ventricular complexes (PVCs), couplets, and runs of NSVT was quantified. Due to the stratified randomization procedure used in the GESICA trial, which was based on the presence or absence of NSVT in the prerandomization Holter, two distinct groups were created; exactly half of the patients received amiodarone in each group. After each clinical visit, at 3, 6, 12, 18, and 24 months, forms containing clinical follow-up data were completed and submitted to the coordinating center. In patients undergoing cardiac surgery or cardiac transplantation, follow-up was discontinued the day in which the surgical procedure was performed. As mentioned, the primary end points of this study were twofold: to determine total mortality and to establish the cause of death as progressive heart failure; sudden death, defined as death within 1 hour of presentation of new symptoms; unknown, defined as the inability to establish with certainty the cause of death; and noncardiac. The principal investigator at each site completed a report detailing the circumstances surrounding the cause of death and, when possible, identifying the immediate cardiovascular cause. Available necropsy reports were sent to the coordinating center. If the cause of death remained uncertain, efforts were made by individual investigators to obtain information from patients' relatives or personal physicians. Individual forms were analyzed by the event committee, who were blinded to patient category, to corroborate and, when indicated, to reclassify the cause of death.
The Kaplan-Meier method was used to construct life-table curves and log-ranks. The percentage reduction in mortality was reported as (1-RR)×100, where RR is the estimated relative risk of an event in the NSVT group compared with the group without NSVT (estimated from the life table). A correction adjustment was done for other variables considered on the likelihood-ratio test on the basis of the proportional hazards model (Cox regression model).
From December 1989 through March 1993, 516 patients were enrolled into the study: 173 (33.5%) in the group with NSVT and 343 (66.5%) in the group without NSVT. Baseline characteristics are depicted in Table 1⇓. At the end of the trial, only 19 patients (3.7%) (5 patients with NSVT and 14 patients without NSVT) were lost to follow-up. These patients were included in the follow-up by using the data from their last completed visit. In 16 patients, follow-up was terminated due to cardiac transplantation (4 patients with NSVT and 12 patients without NSVT). The average follow-up, including patients who died and censored patients, was 13 months (range, 2 to 24 months). The total mortality was 193 of 516 patients (37.4%): 87 of 173 (50.3%) in the group with NSVT and 106 of 343 (30.9%) in the group without NSVT (Table 2⇓).
Table 1⇑ presents a univariate analysis of 30 different clinical, laboratory, ECG, and therapeutic variables at the time of enrollment into the study. Multiple logistic regression analysis identified five independent prerandomization predictors of the presence of NSVT in the 24-hour Holter: increased furosemide dose (P<.0001), decreased systolic blood pressure (P<.0002), increased serum creatinine level (P<.002), faster heart rate (P<.01), and Chagas' disease as the etiology for heart failure (P<.05).
NSVT as a Predictor of Total Mortality
Compared with patients without NSVT, Kaplan-Meier curves depicted higher mortality in patients with NSVT. Intercurve separation became evident during early follow-up and became even more obvious throughout the 2 years of follow-up (Fig 1⇓). NSVT increased total mortality (RR, 1.69; 95% CI, 1.27 to 2.24; P<.0002). Cox regression model for total mortality demonstrated that NSVT maintained an independently increased RR of 1.62 (95% CI, 1.22 to 2.16; P<.001) (Table 2⇑).
NSVT as a Predictor of the Death Mode
An increase in sudden death in the group with NSVT was observed from the beginning of the survival curve (log-rank test P<.001) (Fig 2⇓, left). The RR (2.77; 95% CI, 1.78 to 4.44; P<.001) of sudden death in the group with NSVT was higher than that of total mortality, Cox regression model correction RR (2.56; 95% CI, 1.6 to 4.1) (P<.0001) (Table 2⇑). The increase in total mortality in the group with NSVT (50.29% versus 30.49%) was mainly due to an increase in sudden death (23.7% versus 8.75%) (Table 2⇑). In contrast, a nonsignificantly lower mortality was observed in the group of patients without NSVT due to progressive heart failure (log-rank test P=.22) (Fig 2⇓, right).
NSVT as a Predictor of Increased Total Mortality in Different Subgroups
The presence of NSVT resulted in a similar increase in the RR of total mortality regardless of functional class: class II advanced RR, 1.87 (95% CI, 0.81 to 4.33; P<.064); class III RR, 1.83 (95% CI, 1.21 to 2.77; P<.0016); and class IV RR, 1.44 (95% CI, 0.92 to 2.44; P<.056).
Ischemic and Nonischemic Cardiomyopathy
The diagnosis of ischemic cardiomyopathy was established in 39% of patients with use of the following criteria: past history of acute myocardial infarction accompanied by Q waves in the ECG, history of myocardial revascularization, and/or an angiogram with ≥75% of lumen obstruction in one or more epicardial coronary arteries when available. Interestingly, the incidence of NSVT in the group with ischemic cardiomyopathy was similar to that encountered in the nonischemic group. The increased mortality associated with the presence of NSVT in the ischemic cardiomyopathic group (RR, 1.95; 95% CI, 1.23 to 3.04) also was not statistically different from that of the nonischemic cardiomyopathic group (RR, 1.51; 95% CI, 1.06 to 2.15).
Treatment With Amiodarone
The increased mortality associated with the presence of NSVT was noted regardless of whether the group was (RR, 1.55; 95% CI, 1.03 to 2.39) or was not (RR, 1.82; 95% CI, 1.25 to 2.81) treated with amiodarone.
Predictors of Total Mortality and Sudden Death in the Admission Holter
A quantitative analysis of the 24-hour preenrollment Holter was performed in the first 300 patients recruited into the study. The Holter recordings of 5 patients were excluded from the analysis because the tape contained ≤20 hours of recording. Data were independently reviewed by two observers.
No correlation was found between total mortality and the number of PVCs per hour or the number of episodes of NSVT per patient. Although significance was observed in the univariate analysis in patients with ≥10 PVCs per hour (RR, 1.59; P<.05) and NSVT of ≥4 beats (RR, 1.30; P<.01), both disappeared when the Cox regression model was used. Total mortality analysis documented significance with NSVT (RR, 1.75; 95% CI, 1.23 to 2.50; P<.002) and couplets (RR, 1.81; 95% CI, 1.22 to 2.66; P<.002). The inclusion of couplets and/or NSVT as a risk factor called ventricular repetitive beats (VRBs) resulted in a higher RR (2.20; 95% CI, 1.45 to 3.34; P<.0001).
For sudden death, all RRs increased significantly: NSVT, 2.43 (95% CI, 1.34 to 4.42; P<.005), couplets, 3.37 (95% CI, 1.57 to 7.25; P<.0005), and VRBs, 5.30 (95% CI, 2.09 to 13.4; P<.00005) (Table 3⇓).
No predictive value was found between Holter findings and progressive heart failure death. Cox regression model using clinical and Holter variables denoted for VRBs an RR value of 2.90 for total mortality (95% CI, 1.10 to 7.64; P<.05) and an RR value of 10.1 for sudden death (95% CI, 1.91 to 52.7; P<.01). Accordingly, the sensitivity of sudden death was 89% (39 of 44), with a specificity of 42% (105 of 251). For the subgroup of patients who died suddenly (14.9%; 44 of 295), the positive predictive value was 21.1% with a negative predictive value of 95.5% (false-negative rate, 4.5%).
This multicenter study was an evaluation in a large cohort of patients with severe heart failure and high 2-year mortality (47.2%) of the predictive value of the presence of NSVT and other ventricular arrhythmias diagnosed with 24-hour Holter monitoring. The results clarified the controversial issue of whether NSVT is a predictor of high risk for sudden death6 10 11 versus solely an additional marker of the severity of heart failure and, as such, a harbinger of mortality due to progressive heart failure.5 12 15 Our results are in keeping with other studies12 13 14 15 16 17 18 19 that concluded that the presence of NSVT correlates with total mortality, with a persistent increased risk of 1.63 after adjustment with other variables.
Sudden death associated with asymptomatic ventricular arrhythmias is frequently observed in patients with chronic heart failure. However, previous investigations that have attempted to address this relation have been retrospective and involved only small populations. The increased total mortality observed in our patients with NSVT was mainly due to the occurrence of a higher incidence of sudden death in this latter group. In contrast, no intergroup differences were noted when death occurred secondary to progression of heart failure. The RR for sudden death in patients with NSVT (RR, 2.77) was higher than that calculated for total mortality.
Although the GESICA trial14 was a study of the effect of low-dose amiodarone in patients with severe heart failure, analysis of the predictive value of NSVT on total mortality was a prospective secondary objective of that study. Because our operant hypothesis was that total mortality would be reduced via a reduction in sudden death in the group of patients treated with amiodarone, a stratified randomization procedure was performed according to the presence of NSVT. Treatment with amiodarone did not confound the results of this study because the relatively higher risk of death associated with the presence of NSVT persisted despite antiarrhythmic therapy. A trend toward reduced mortality was observed in the patients enrolled in the GESICA trial with or without NSVT, with borderline significance noted in patients with NSVT—probably due to the higher mortality noted in this patient subset. In addition, the reduction in mortality was due to the lower incidence of both sudden death and death as a result of progression of heart failure.
Despite the relative homogeneity of the population that was studied, patients with NSVT had a greater impairment of ventricular function. NSVT was also associated with higher doses of diuretic therapy, lower systolic blood pressure, higher creatinine levels, and faster heart rates, all of which tend to indicate an association between the presence of NSVT and worsening of clinical status. Interestingly, the increases in the RR of death were noted to be similar regardless of the functional class. Accordingly, patients with heart failure and NSVT appear to have a similar absolute mortality to that of a patient in a worse functional class but without NSVT.
The range in which the incidence of NSVT has been reported in heart failure is 28% to 80%.17 The relatively low incidence of 33.4% observed in our study is probably due to strict adherence to the exclusion criteria. Also, therapy with converting enzyme inhibitors in >90% of patients might have influenced the prevalence of NSVT.18
The correct definition of sudden death in highly symptomatic patients with heart failure remains ill defined. Accordingly, discrepancies in the definition of sudden death might account, in part, for the differences in incidence between different studies.20 In the present study, a simple, yet strict, definition of sudden death was used. Sudden death, often called “arrhythmic death,” has been found to be secondary to sustained ventricular tachycardia and/or ventricular fibrillation in >80% of the patients who have died at a time when an ambulatory Holter monitor study was being recorded.7 Although most studies have documented bradyarrhythmias in the remaining 20% of patients as the cause of sudden death, Luu et al21 described this mechanism in >60% of sudden deaths occurring in patients with severe heart failure. However, these data should be viewed cautiously because they were obtained in a retrospective fashion from hospitalized patients awaiting heart transplantation.
A high proportion of patients included in the GESICA trial had a nonischemic etiology for their heart failure. This did not appear to influence the results of that study because no etiology-related differences were noticeable.
In the evaluation of the quantitative Holter sample studied, the presence of couplets showed a greater independent prognostic value than NSVT for the incidence of sudden death. Couplets were found in 59% of the population and in >90% of the patients with NSVT. Couplets and/or NSVT (VRBs) were detected in 62.7% of the study population, with a 50.8% mortality rate. The remaining 37.3% without VRBs had a lower mortality rate, 26.3%.
Consequently, the presence of VRBs constituted a significant marker of sudden death, with a sensitivity and specificity of 89% and 42%, respectively. Our Holter sample had a 15% incidence of sudden death, with a positive VRBs predictive value of only 21%. However, the absence of VRBs remarkably decreased the incidence of sudden death to 4.5% (negative predictive value, 95.5%). In contrast to other studies,10 we did not find a correlation between the number and length of NSVT episodes and sudden death. To date, there is no evidence to support the fact that antiarrhythmic suppression of complex ventricular arrhythmias in patients with severe heart failure can reduce the incidence of total or sudden death. Furthermore, suppression of ventricular arrhythmias may be a marker for lower risk of sudden death even without antiarrhythmic treatment during follow-up.22 23 The latter could have accounted for the unusually low mortality rate observed in the placebo group of the CAST study,24 in which patients were admitted according to suppression criteria with antiarrhythmic drugs.
Accordingly, pharmacological suppression of arrhythmias should not be considered a surrogate end point for lowering the incidence of sudden death because estimation of complex ventricular arrhythmias as risk markers does not imply a direct cause-effect relation. This direct association would represent only one of the possible explanations, considering the complex causality of rhythm disturbances.
In summary, detection of NSVT and VRBs enables the stratification of sudden death in patients with advanced heart failure. Furthermore, the absence of the latter in a 24-hour Holter is of great clinical relevance because it may indicate a low probability of sudden death.
Appendix: GESICA-GEMA Centers and Investigators
Instituto del Corazo´n del Hospital Italiano (Buenos Aires, Argentina): Herna´n Doval, Rodolfo Curiel, Miguel Sedano y Varela; Sanatorio Guemes and Fundacio´n Favaloro (Buenos Aires): Sergio Perrone, Carlos Campanini, Juan Delgado, and Pablo Comignani; Instituto de Cardiologi´a del Hospital Espan˜ol (Buenos Aires): Hugo Grancelli, Guillermo Bortman, Jose´ Buscema, and Victorio Carosella; Hospital Municipal Argerich (Buenos Aires): Daniel Nul, Lila Leyro, and Ruben Kevorkian; Hospital Municipal Santojanni (Buenos Aires): Noemi´ Prieto and Hugo Carreau; Hospital Israelita (Buenos Aires): Saul Soifer, Hector Cercos, and Hugo Torres; Instituto Anta´rtida (Buenos Aires): Sergio Varini, Mari´a Ramos, and Viviana Perujini; Hospital Cullen (Santa Fe´): Carlos Becker and Mario Sejas; Hospital Aeronau´tico (Buenos Aires): Gustavo Cerezo; Hospital Padilla (Tucuma´n): Sergio Haguad; Hospital Alema´n (Buenos Aires): Jose´ Gant Lopez; Instituto Pombo (Buenos Aires): Humberto Gugliotta; Hospital Municipal Fernandez (Buenos Aires): Daniel Agranatti; Hospital Italiano (La Plata, Pcia Buenos Aires): Gustavo Vigo; Cli´nica Mari´a Auxiliadora (Olavarria, Pcia Buenos Aires): Roberto Balado and Ernesto Ylarri; Instituto Cardiovascular (Buenos Aires): Marcelo Trivi; Hospital Araoz Alfaro (Pcia Buenos Aires): Susana Fernandez; Hospital Irurzun (Queque´n, Pcia Buenos Aires): Gustavo Casas; Cli´nica Bazterrica (Buenos Aires): Adrian Charak; Hospital Municipal Pirovano (Buenos Aires): Andre´s Graziano; Policli´nico Ados (Neuque´n): Carlos Lavergne; Sanatorio Anchorena (Buenos Aires): Mari´a Sosa Liprandi; Hospital San Marti´n (Corrientes): Julio Iban˜ez; Sanatorio Belgrano (Mar del Plata, Pcia Buenos Aires): Luis Moreno; Sanatorio Municipal Julio Mendez (Buenos Aires): Eda Monetti; and Hospital Municipal Durand (Buenos Aires): Alejandro Hershon.
We thank the medical investigators who participated in this trial, Adriana Grosman for translation of the manuscript, and Roemmers Argentina Laboratory for its cooperation.
*Group members and institutional affiliations are listed in “Appendix.”
- Received January 25, 1996.
- Revision received July 24, 1996.
- Accepted July 31, 1996.
- Copyright © 1996 by American Heart Association
Dargie H, Cleland J, Leckie B, Inglis C, East B, Ford I. Relation of arrhythmias and electrolyte abnormalities to survival in patients with severe chronic heart failure. Circulation. 1987;75(suppl IV):IV-98. Abstract.
Nickolic G, Bishop R, Singh J. Sudden death recorded during Holter monitoring. Circulation. 1982;66:218-225.
Francis G. Development of arrhythmia in the patient with congestive heart failure: pathophysiology, prevalence and prognosis. Am J Cardiol. 1986;57(suppl):3B-7B.
Massie B, Francis G, Tandon P, Anderson S, De Mets D, Packer M, on behalf of the PROMISE Investigators. Asymptomatic ventricular arrhythmias do not identify patients with severe heart failure at risk for sudden death: results of the PROMISE trial. J Am Coll Cardiol. 1993;21:459A. Abstract.
Doval H, Nul D, Grancelli H, Perrone S, Bortman G, Curiel R. Randomised trial of low-dose amiodarone in severe congestive heart failure. Lancet. 1994;1:493-498.
Packer M. Lack of relation between ventricular arrhythmias and sudden death in patients with chronic heart failure. Circulation. 1992;85(suppl I):I-50-I-56.
Bigger T. Why patients with congestive heart failure die: arrhythmias and sudden cardiac death. Circulation. 1987;75(suppl IV):IV-28-IV-35.
Fletcher R, Cintron G, Johnson G, Orndorff J, Carson P, Cohn J, for the V-HeFT II VA Cooperative Studies Group. Enalapril decreases prevalence of ventricular tachycardia in patients with chronic congestive heart failure. Circulation. 1993;87(suppl VI):VI-49-VI-55.
Cohn J, Johnson G, Shabetai R, Loeb H, Tristani F, Rector T, Smith R, Fletcher R, for the V-HeFT VA Cooperative Studies Group. Ejection fraction, peak exercise oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation. 1993;87(suppl VI):VI-5-VI-16.
Luu M, Stevenson W, Stevenson L, Baron K, Walden J. Diverse mechanisms of unexpected cardiac arrest in advanced heart failure. Circulation. 1989;80:1675-1680.
Goldstein S, Brooks M, Ledingham R, Kennedy H, Epstein A, Pawitan Y, Bigger J. Association between ease of suppression of ventricular arrhythmia and survival. Circulation. 1995;91:79-83.
Pratt C, Moye´ L. The Cardiac Arrhythmias Suppression Trial: casting suppression in a different light. Circulation. 1995;91:245-247.