Author + information
- Received October 17, 1996
- Revision received January 31, 1997
- Accepted February 3, 1997
- Published online May 1, 1997.
- Daniel R. Nul, MDA,*,
- Hernán C. Doval, MDA,
- Hugo O. Grancelli, MDA,
- Sergio D. Varini, MDA,
- Saul Soifer, MDA,
- Sergio V. Perrone, MDA,
- Noemí Prieto, MDA,
- Omar Scapin, MDA,
- on Behalf of the GESICA-GEMA Investigators1
- ↵*Dr. Daniel R. Nul, GESICA Group, Araoz 788, Buenos Aires, Argentina, CP 1414.
Objectives. The impact of amiodarone on mortality in patients with severe congestive heart failure (CHF) (New York Heart Association functional classes II [advanced], III and IV; left ventricular ejection fraction <35%) in the Grupo de Estudio de la Sobrevida en la Insuficiencia Cardiaca en Argentina (GESICA) trial was analyzed in relation to initial mean baseline heart rate (BHR) and its change after 6 months of follow-up.
Background. Trials of amiodarone therapy in CHF have produced discordant results, suggesting that the effect is not uniform in all patient subgroups with regard to survival.
Methods. The present analysis was carried out in 516 patients randomized to receive amiodarone, 300 mg/day (n = 260), or nonantiarrhythmic therapy (n = 256, control group) and followed up for 2 years. Survival was evaluated for patients with a BHR ≥90 beats/min (control: n = 132; amiodarone: n = 122) and <90 beats/min (control: n = 124; amiodarone: n = 138). Survival was also analyzed according to heart rate reduction at 6 months for 367 patients.
Results. For patients with a BHR ≥90 beats/min, amiodarone therapy reduced mortality to 38.4% compared with 62.4% in control patients (relative risk [RR] 0.55, 95% confidence interval [CI] 0.35 to 0.95, p < 0.002). Both sudden death (RR 0.46, 95% CI 0.24 to 0.90, p < 0.02) and progressive heart failure death (RR 0.60, 95% CI 0.30 to 1.03, p < 0.06) were reduced, and functional capacity was improved. In patients with a BHR <90 beats/min, amiodarone did not alter survival. Among 367 patients who completed 6 months of follow-up, amiodarone reduced 2-year mortality only in those with a BHR ≥90 beats/min, which was reduced at 6 months.
Conclusions. Elevated rest heart rates in severe CHF identify a subgroup of patients who benefit from treatment with amiodarone. Amiodarone-induced heart rate slowing may be an important benefit for patients.
(J Am Coll Cardiol 1997;29:1199–205)
The prevalence of ventricular arrhythmias and sudden death in severe congestive heart failure (CHF) has focused attention on antiarrhythmic therapy as a means for reducing mortality ([1–5]). Amiodarone has emerged as the drug of choice because of its potent antiarrhythmic effects and low potential for proarrhythmic and negative inotropic effects, as opposed to other antiarrhythmic agents ([6–9]). Grupo de Estudio de la Sobrevida en la Insuficiecia Cardiaca en Argentina (GESICA), a randomized trial of low dose amiodarone in severe CHF reported a 28% reduction in 2-year overall mortality (). The mortality reduction was due not only to a decrease in sudden death, but also to a reduction in death from progressive heart failure. The reduction in mortality was independent of the severity of ambient ventricular arrhythmia on admission, and was associated with a decrease in hospital admissions due to heart failure and an improvement of functional capacity in amiodarone-treated patients. The results suggest benefit beyond that expected from purely an antiarrhythmic effect.
In addition to its antiarrhythmic action, amiodarone slows the sinus heart rate. The precise cause of the reduction in rate is unclear, but it may be due to depression of sinus node automaticity and slowing of atrioventricular node conduction (in the case of atrial fibrillation) mediated in part by noncompetitive beta-blockade ([11, 12]).
An inappropriate, rapid heart rate in CHF may reflect abnormal activation of the sympathetic nervous system, which might be reduced by amiodarone ([13, 14]). We hypothesized that the benefit of amiodarone may be mediated in part by a reduction in heart rate. The objective of this report was to retrospectively analyze the mortality reduction produced by amiodarone in GESICA, relative to the baseline heart rate (BHR) and its reduction during treatment.
1.1 Study design and patient selection
The design of the GESICA trial and the characteristics of the patients enrolled have been previously published elsewhere and will be briefly summarized (). The present analysis was carried out in the entire study group of 516 patients. All patients had advanced CHF with a marked reduction of left ventricular systolic function that was adequately treated with a low sodium diet, diuretics, digitalis and angiotensin-converting enzyme inhibitors, without antiarrhythmic treatment. Patients were required to have stable functional capacity judged by a combination of the Canadian Cardiovascular Society criteria and the Specific Activity Scale, as equivalent to New York Heart Association functional class II (advanced), III or IV.
Also, at least two of the three following indexes of systolic myocardial dysfunction had to be present: chest radiograph cardiothoracic ratio >0.55, ejection fraction measured by radioisotope method ≤0.35 and end-diastolic echocardiographic diameter ≥3.2 cm/m2.
Exclusion criteria were amiodarone treatment during the preceding 3 months; thyroid dysfunction; severe respiratory failure; concomitant serious disease; mitral stenosis, aortic stenosis and/or hypertrophic or restrictive cardiomyopathy; recent instability with angina, myocardial infarction, heart failure onset or syncope within the previous 3 months; atrioventricular conduction disorder; history of sustained ventricular tachycardia or ventricular fibrillation; and asymptomatic ventricular tachycardia for >10 beats with heart rate faster than 100 beats/min.
The study was approved by the Institutional Review Board of each hospital, and each patient gave informed consent. An independent scientific and ethics committee monitored the progress of the study.
The study was conducted in 26 hospitals in Argentina from December 1989 to March 1993. This was a prospective, parallel, randomized and stratified trial according to the presence of nonsustained ventricular tachycardia in the initial 24-h Holter recording.
The control group (n = 256) continued to receive nonantiarrhythmic therapy for heart failure. The amiodarone group (n = 260) received amiodarone at a dose of 600 mg/day for 14 days, and then 300 mg/day was added to their standard heart failure therapy. Patients were seen at 3, 6, 12, 18 and 24 months for follow-up. When a patient required heart surgery or transplantation, follow-up was discontinued on the day of the surgical procedure.
End points were total mortality, cardiopulmonary resuscitation and symptomatic sustained ventricular tachycardia. Death was attributed to progressive heart failure, sudden death (death within an hour of presentation of new symptoms), unknown (no means of establishing with certainty the cause of death) and noncardiac causes. The cause of death was determined by investigators at the coordinating center who had no knowledge of the assigned treatment group. The study was terminated by the steering committee, which acted on the advice of the scientific and ethics committee, after the second prespecified analysis.
1.2 Analysis according to heart rate
Heart rate was determined from a physical examination. In the presence of atrial fibrillation, the average heart rate for 1 min was assessed. The effect of BHR was evaluated by stratifying each group (amiodarone and control) according to a mean BHR ≥90 and <90 beats/min, respectively. Thus, four groups were established, each comprising approximately the same number of patients.
In addition, 367 patients completed 6 months of follow-up, at which time rest heart rate was re-evaluated. For the purposes of analysis, those patients whose rest heart rate at 6 months of follow-up was less than the BHR were considered to have had a reduction in heart rate.
1.3 Statistical analysis
The Kaplan-Meier method was used to construct life-table curves, and the curves were compared using a two-tailed log-rank test.
Percent reduction in mortality was reported as: (1 − RR) × 100, where RR is the estimated relative risk of an event in the compared groups. To adjust for baseline differences, the proportional hazards model (Cox regression model) was used, incorporating significant variables (p < 0.05) according to the likelihood ratio test.
To compare differences between groups, the Student ttest was used for evaluation of continuous variables and the chi-square test was used for proportions.
Follow-up, including death and censored patients, ranged from 2 to 24 months (average 13). The 2-year actuarial mortality rate was 48%. At the end of the trial, 19 patients (3.7%) had been lost to follow-up and were censored from the analysis on the last known date of follow-up: 9 patients with a BHR <90 beats/min (5 in control group and 4 in amiodarone group) and 10 with a BHR ≥90 beats/min (7 in control group and 3 in amiodarone group). In 16 patients, follow-up was concluded at the time of heart transplantation: 5 patients with a BHR <90 beats/min (2 in control group and 3 in amiodarone group) and 11 patients with a BHR ≥90 beats/min (5 in control group and 6 in amiodarone group). In six patients, follow-up was concluded at the time of other cardiovascular surgery: three patients with a BHR <90 beats/min (all in control group) and 3 patients with a BHR ≥90 beats/min.
2.1 Clinical characteristics and baseline heart rate
From the total group of 516 patients, 254 had a BHR ≥90 beats/min (control: n = 132; amiodarone: n = 122) and 262 patients had a BHR <90 beats/min (control: n = 124; amiodarone: n = 138).
Baseline clinical characteristics differed among the groups (Table 1). Patients with a BHR ≥90 beats/min had evidence of worse heart failure compared with patients with a BHR <90 beats/min, as indicated by a lower left ventricular ejection fraction (17.5% vs. 20.3%, p < 0.003), lower systolic blood pressure (p < 0.005), lower serum sodium concentration (p < 0.003), worse functional class (p < 0.001), greater evidence of right heart failure (p < 0.0001) and higher diuretic requirement (p < 0.003).
However, there were no differences between control and amiodarone-treated patients in the group with a BHR >90 beats/min (Table 2) or in the group with a BHR <90 beats/min.
2.2 Baseline heart rate, mortality and amiodarone treatment
For the entire study group, the 2-year mortality rate was higher in patients with a BHR ≥90 beats/min (50.8%) compared with patients with a BHR <90 beats/min (44.8%) (p < 0.01). For patients with a BHR ≥90 beats/min, the mortality rate was lower in amiodarone-treated patients compared with control patients (38.4% vs. 62.4%, RR 0.55, 95% confidence interval [CI] 0.35 to 0.95, p < 0.002) (Table 3, Fig. 1).
In patients with a BHR <90 beats/min, amiodarone-treated patients had a mortality rate similar to that of the control patients (44.8% vs. 44.7%, RR 1.0, 95% CI 0.74 to 1.45, p < 0.97) (Table 3, Fig. 1).
Patient division in quintiles according to BHR indicates that mortality reduction with amiodarone was associated with a higher BHR, being highly significant in the higher quintile with a BHR ≥100 beats/min (RR 0.35, 95% CI 0.19 to 0.67, p < 0.001). At the opposite end, in the lowest quintile with a BHR <76 beats/min, a trend toward an increase in risk in amiodarone-treated patients was observed (RR 1.34, 95% CI 0.68 to 2.67, p < 0.35) (Fig. 2).
2.3 Baseline heart rate, amiodarone treatment and mode of death
A significant reduction in sudden death was observed in patients with a BHR ≥90 beats/min treated with amiodarone (RR 0.46, 95% CI 0.24 to 0.90, p < 0.02). Death due to progressive heart failure was also lower in this group of patients, but this difference was of borderline statistical significance (RR 0.60, 95% CI 0.30 to 1.03, p < 0.06). In patients with a BHR <90 beats/min, no significant differences in sudden death or death from progressive heart failure were observed (Table 3).
2.4 Heart rate modification
A total of 367 patients had their rest heart rate determined at 6 months of follow-up. The BHR at the beginning of the study was similar in the amiodarone and control groups. In those patients with a BHR ≥90 beats/min (188 patients, mean heart rate 100.9 ± 10 beats/min), heart rate was reduced at 6 months by 12.7 ± 16 beats/min in the amiodarone group and by 7.6 ± 13 beats/min in the control group (p < 0.02). In 179 patients with a BHR <90 beats/min (mean heart rate 76.7 ± 7 beats/min), there was a trend toward an increased heart rate at 6 months, which was greater in the control group (6.8 ± 12 beats/min) than in the amiodarone group (2.2 ± 12 beats/min) (p < 0.01). There were no differences between the groups with regard to medication changes over follow-up. Amiodarone withdrawal occurred in only 15 patients.
2.5 Heart rate reduction and mortality
Of the 137 patients with a BHR ≥90 beats/min and whose heart rate decreased at 6 months of follow-up, the mortality rate was significantly lower in those patients treated with amiodarone (21.7%) compared with control patients (53.8%) (p < 0.002). In patients with a BHR ≥90 beats/min and whose heart rate was not decreased at 6 months of follow-up, the mortality rate was similar for amiodarone-treated (55.2%) and control patients (55.9%). Among patients with a BHR <90 beats/min, there were no differences in mortality, regardless of changes in heart rate at 6 months of follow-up.
2.6 Functional capacity modifications
Baseline functional capacity was similar in the amiodarone and control groups. In patients with a BHR ≥90 beats/min, the proportion of patients in functional class I or II at 6 months was greater in the amiodarone group than in the control group (51.3% vs. 31.5%), and the proportion of patients in functional class IV was lower in the amiodarone group than in the control group (17.8% vs. 33.7%) (p < 0.02). In patients with a BHR <90 beats/min, no differences were observed in functional capacity at 6 months between the amiodarone and control groups.
In the present report the addition of low dose amiodarone to conventional therapy in severe CHF decreased mortality, but only among patients with a high rest heart rate at baseline. Amiodarone strikingly reduced 2-year overall mortality in patients with a BHR ≥90 beats/min, but had no impact on survival in patients with a BHR <90 beats/min. The beneficial effect was due to a decrease in both sudden death and death from progressive heart failure. Furthermore, for patients surviving for 6 months, the benefit in the 2-year survival rate and improved functional class was confined to those whose heart rate was reduced during follow-up.
3.1 Baseline heart rate and mortality
A rapid heart rate is associated with a variety of prognostic factors indicative of worse heart failure and greater mortality. Interestingly, in the patients treated with amiodarone, this was not the case; mortality was similar across all BHR ranges. Therefore, the higher the initial rest heart rate, the greater the benefit observed from therapy with amiodarone (Fig. 2). In the highest quintile with a BHR ≥100 beats/min, the mortality risk reduction was 65%, with the narrowest CI. As previously reported, in the entire GESICA trial group, there was a favorable trend toward a reduction in sudden and progressive heart failure deaths, which did not attain statistical significance (). In the present report reductions in these end points were greater and achieved statistical significance in patients with a BHR ≥90 beats/min. A tendency toward increased mortality in the lowest quintile with a BHR <76 beats/min (Fig. 2) could suggest that bradyarrhythmias were increased in this group, contributing to sudden death. This seems unlikely, however, because the mode of death (sudden vs. progressive heart failure) in patients with a BHR <76 beats/min was similar in those treated with amiodarone compared with control patients.
3.2 Amiodarone’s effect in congestive heart failure
Amiodarone is a unique and complex drug with multiple effects. It is difficult to determine which of its effects might be related to our findings. In contrast to many antiarrhythmic drugs, amiodarone has been effective in several patient groups and has not been associated with increased mortality (). Its most obvious direct electrophysiologic effect is prolongation of the action potential duration and repolarization time (class III antiarrhythmic effect). Despite an association of QT prolongation with malignant ventricular arrhythmias, amiodarone is the antiarrhythmic agent that has the least proarrhythmic effect, possibly due to more uniform prolongation of myocardial repolarization resulting in a decrease of QT dispersion ([15–18]). Sudden death reduction in patients with high heart rates may be explained by an antiarrhythmic or antifibrillatory effect. However, an antiarrhythmic effect does not account for all of our findings.
In the present study functional capacity improved, hospital admissions for heart failure were reduced and mortality due to progressive heart failure was diminished. This can be connected to several studies showing that therapy, even with lower amiodarone dose, increases left ventricular ejection fraction ([19–22]).
In another large amiodarone trial involving milder heart failure Congestive Heart Failure—Survival Trial of Antiarrhythmic Therapy (CHF-STAT), although neutral survival results were reported, a remarkable finding was that ejection fraction increased by 42% at 6 months and that improvement was maintained at 2 years (). The highest increase in ejection fraction was reported in patients with the greatest heart rate reduction, who showed a favorable survival trend (). That trend is more significant in patients with nonischemic causes of heart failure. Although randomization was not balanced for etiology in the GESICA trial, results were similar anyway, and it does not explain contrasting results between both trials ().
The present analysis provides further evidence that differences in the severity of heart failure are a major reason for the discrepant findings of the CHF-STAT and GESICA studies. Patients in GESICA had more severe CHF with an average BHR of 90 beats/min for the entire study group, compared with 80 beats/min in CHF-STAT. Accordingly, if only the patients with a high BHR and heart rate reduction during treatment benefit, the neutral results in the CHF-STAT would be expected. Heart rate reduction with the amiodarone dose used in the GESICA trial was observed at 3 months of therapy, attained a steady state after 6 months and was maintained during the remainder of follow-up. Patients with a higher BHR showed a more dramatic heart rate reduction during amiodarone treatment. A slighter decrease in mean heart rate was observed in the control group with a BHR ≥90 beats/min and might be interpreted as an indication of improved clinical condition in survivors after 6 months.
3.3 Heart rate reduction hypothesis
The precise nature of amiodarone-induced bradycardia remains unclear, but it may result from direct depression of sinus node automaticity and from noncompetitive antiadrenergic action (). The antiadrenergic effect of amiodarone has been suggested to result from inhibition of adenylate cyclase formation and a reduction in the number of beta-adrenergic receptors, an effect that becomes more prominent after long-term amiodarone treatment (). This raises the possibility that beneficial effects could be mediated by prevention of catecholamine toxicity and improved myocardial energetics, as has been suggested for therapy with beta-adrenergic blocking agents in heart failure ([24, 25]). These effects are more likely to have an impact on patients with severe CHF who have greater sympathetic activation, higher heart rates and the most dramatic heart rate reduction during treatment. In the Cardiac Insufficiency BIsoprolol Study (CIBIS), beneficial effects included an increase in ejection fraction, a reduction in heart rate, a decrease in hospital admissions due to heart failure and a beneficial trend toward improved survival in nonischemic patients (). Recently, pooled results from Carvedilol trials reported a marked reduction in mortality with the use of this beta-blocker in patients with CHF (). Interestingly, mortality reduction was clearer in patients with a BHR >82 beats/min (RR 0.26, 95% CI 0.12 to 0.55) and was not related to the etiology of heart failure. These trials include patients with less severe heart failure than in GESICA, in which beta-adrenergic blocking therapy would have been difficult to initiate because of the severity of heart failure.
Several agents used for the treatment of heart failure have been shown to increase ejection fraction and functional capacity, but this did not necessarily improve survival ([28–30]). It is possible that heart rate reduction by itself is responsible for the improvement in ventricular function with amiodarone ([14, 24, 25]). An abnormal force–frequency relation in failing myocardium results in further depression of contractility at fast heart rates (). Amiodarone-induced heart rate slowing may improve cardiac function, and prolongation of the action potential duration may also have an inotropic effect owing to greater intracellular calcium availability. A longer diastolic filling period also promotes greater coronary blood flow, which is likely to reduce ischemia. Amiodarone also has vasodilating properties that could be beneficial.
3.4 Clinical implications
Our analysis was retrospective. It may be helpful in future heart failure studies with amiodarone or heart rate–reducing agents to stratify the randomization according to heart rate to further address this issue.
These results have important clinical implications. We suggest that chronic amiodarone therapy be recommended for patients with CHF who have high rest heart rates and avoided for those with slower rates. Furthermore, because many of amiodarone’s side effects are dose related, heart rate reduction can potentially be used as a guide to adjust the amiodarone dose to obtain the best risk–benefit relation for improving heart failure and reducing mortality.
We thank the medical investigators who participated in this trial: Adriana Grosman for translation of the manuscript; William Stevenson, MD, for revision of the manuscript; and Roemmers Argentina Laboratory, Buenos Aires, Argentina for its cooperation.
A.1 GESICA-GEMA Centers and Investigators
Instituto del Corazón del Hospital Italiano (Buenos Aires): Hernán C. Doval, MD, Rodolfo Curiel, MD, Miguel Sedano y Varela, MD. Sanatorio Guemes and Fundación Favaloro (Buenos Aires): Sergio V. Perrone, MD, Carlos Campanini, MD, Juan Delgado, MD, Pablo Comignani, MD. Instituto de Cardiologı́a del Hospital Español (Buenos Aires): Hugo O. Grancelli, MD, Guillermo Bortman, MD, José Buscema, MD, Victorio Carosella, MD. Hospital Municipal Argerich (Buenos Aires): Daniel R. Nul, MD, Lila Leyro, MD, Ruben Kevorkian, MD. Hospital Municipal Santojanni (Buenos Aires): Noemı́ Prieto, MD, Hugo Carreau, MD. Hospital Israelita (Buenos Aires): Saul Soifer, MD, Hector Cercos, MD, Hugo Torres, MD. Instituto Antártida (Buenos Aires): Sergio Varini, MD, Marı́a Ramos, MD, Viviana Perujini, MD. Hospital Cullen (Santa Fé): Carlos Becker, MD, Mario Sejas, MD. Hospital Aeronaútico (Buenos Aires): Gustavo Cerezo, MD. Hospital Padilla (Tucumán): Sergio Haguad, MD. Hospital Alemán (Buenos Aires): José Gant Lopez, MD. Instituto Pombo (Buenos Aires): Humberto Gugliotta, MD. Hospital Municipal Fernandez (Buenos Aires): Daniel Agranatti, MD. Hospital Italiano (La Plata, Pcia Buenos Aires): Gustavo Vigo, MD. Clinica Maria Auxiliadora (Olavarria, Pcia Buenos Aires): Roberto Balado, MD, Ernesto Ylarri, MD. Instituto Cardiovascular (Buenos Aires): Marcelo Trivi, MD. Hospital Araoz Alfaro (Pcia Buenos Aires): Susana Fernandez, MD. Hospital Irurzun (Quequén, Pcia Buenos Aires): Gustavo Casas, MD. Clinica Bazterrica (Buenos Aires): Adrian Charak, MD. Hospital Municipal Pirovano (Buenos Aires): Andrés Graziano, MD. Policlı́nico Ados (Neuquén): Carlos Lavergne, MD. Sanatorio Anchorena (Buenos Aires): Marı́a Sosa Liprandi, MD. Hospital San Martin (Corrientes): Julio Ibañez, MD. Sanatorio Belgrano (Mar del Plata, Pcia Buenos Aires): Luis Moreno, MD. Sanatorio Municipal Julio Mendez (Buenos Aires): Eda Monetti, MD. Hospital Municipal Durand (Buenos Aires): Alejandro Hershon, MD.
Ethics Committee: Mauricio Rosembaum, MD, Carlos Bertolasi, MD, Raul Olivieri, MD.
↵1 The study authors are members of the GESICA Steering Committee and Subcommittees; a complete list of the participating investigators and institutions for the Grupo de Estudio de la Sobrevida en la Insuficiencia Cardiaca en Argentina (GESICA)–Grupo de Estudios Multicéntricos en Argentina (GEMA) trial appears in the AppendixAppendix A.
- baseline heart rate
- congestive heart failure
- Congestive Heart Failure—Survival Trial of Antiarrhythmic Therapy
- confidence interval
- Cardiac Insufficiency BIsoprolol Study
- Grupo de Estudio de la Sobrevida en la Insuficiencia Cardiaca en Argentina
- relative risk
- Received October 17, 1996.
- Revision received January 31, 1997.
- Accepted February 3, 1997.
- The American College of Cardiology
- for the Captopril-Digoxin Study Group,
- Gradman A,
- Deedwania P,
- Cody R,
- et al.
- Bigger T
- Dargie H,
- Cleland J,
- Leckie B,
- Inglis C,
- East B,
- Ford I
- Chatterjee K
- Sager PT,
- Uppal P,
- Follmer C,
- Antimisisiaris M,
- Pruitt C,
- Singh BN
- for the V-HeFT VA Cooperative Studies Group,
- Cohn J,
- Johnson G,
- Shabetai R,
- et al.
- Singh BN,
- Courtney K
- Colatsky TJ,
- Follmer CH,
- Starmer CF
- Hamer AWF,
- Arkles LB,
- Johns JA
- for the CHF-STAT investigators,
- Massie BM,
- Fisher SG,
- Deedwania PC,
- Singh BN,
- Fletcher RD,
- Singh SN
- Anderson B,
- Lomsky M,
- Waagstein F
- CIBIS Investigators and Committees
- Mullieri LA,
- Hasenfuss G,
- Leavitt B,
- Allen PD,
- Alpert NR