Amiodarone in AF With and Without LV Dysfunction
Amiodarone in AF With and Without LV Dysfunction
A total of 3,307 patients enrolled in AFFIRM and AF-CHF trials met inclusion criteria (Fig. 1). Overall, 2,200 patients were randomized to rate control and 1,107 to rhythm control with amiodarone as the first antiarrhythmic agent. Baseline characteristics are summarized in Table 1. The average age was 68.0 ± 0.2 years and 1,027 (31.1%) patients were female. Coronary artery disease was prevalent in 1,441 (43.6%), diabetes in 681 (20.6%), and hypertension in 2,068 (62.5%) patients.
(Enlarge Image)
Figure 1.
Selection of study population from AFFIRM and AF-CHF trials. LVEF = left ventricular ejection fraction.
For patients with an LVEF <30% (N = 842), no difference in baseline characteristics were observed according to whether they were randomized to rate control (N = 476) or rhythm control with amiodarone (N = 366). In contrast, for the subgroup with an LVEF ≥30% (N = 2,483), patients treated with amiodarone (N = 741) were more likely than those randomized to rate control (N = 1,724) to be male, have a higher NYHA functional class, coronary artery disease, and receive β-blockers, lipid lowering drugs, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers.
Over the course of the study, 274 of 1,107 (24.7%) patients initially assigned to amiodarone discontinued the drug. Reasons for discontinuation are listed in Table 2 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9. Recurrence of AF was assessed in eligible patients randomized to rhythm control who received amiodarone as the first antiarrhythmic agent (N = 713; Fig. 1). Baseline characteristics of this patient subgroup are provided in Table 2 according to whether the LVEF was <30% (N = 237), 30–49% (N = 244), or ≥50% (N = 232). Patients with lower LVEFs were younger, more likely to be male, and had a higher prevalence of underlying heart disease and associated comorbidities.
Over an average follow-up of 40.0 ± 0.3 months, 272 (38.1%) patients experienced recurrent AF. Corresponding recurrence-free survival rates were 84%, 72%, and 45% at 1, 2, and 5 years, respectively. Kaplan–Meier AF-free survival curves are shown in Figure 2. No differences in rates of recurrent AF were observed according to LVEF in univariate (P = 0.6306) or multivariate (P = 0.8754) analyses. Univariate and multivariate predictors of recurrent AF are summarized in Table 3, along with nonsignificant selected variables of interest. The only variables independently associated with recurrent AF were diabetes (hazard ratio 1.60, 95% confidence interval CI: [1.15–2.17], P = 0.0044 and larger left atrial size (hazard ratio 1.03 per 1 mm increase, 95% CI [1.01–1.05], P = 0.0096.
(Enlarge Image)
Figure 2.
Freedom from recurrent atrial fibrillation according to left ventricular systolic function. Shown are Kaplan–Meier curves depicting freedom from recurrent atrial fibrillation in the 713 patients randomized to rhythm control, in sinus rhythm at time 0, and receiving amiodarone as the first-line antiarrhythmic agent. No statistically significant difference in freedom from recurrent atrial fibrillation was observed according to left ventricular ejection fraction (LVEF) stratified as ≥50%, 30–49%, and <30%.
Patients who were excluded from the analysis of AF recurrence by virtue of failing to achieve normal sinus rhythm at time zero (N = 345) had a similar all-cause mortality rate to those in sinus rhythm (adjusted P = 0.6875) but a higher all-cause hospitalization rate (adjusted P < 0.0001). The lack of association between LVEF and AF recurrence was corroborated by a sensitivity analysis that included the 345 patients in AF at the 4-month visit by considering AF to have recurred at day 1 (P = 0.2075). Similarly, forcing the variable "rhythm at the time of echocardiography" into multivariate models had no appreciable effect on outcomes.
The proportion of time spent in AF was significantly higher in patients randomized to rate control therapy compared to rhythm control with amiodarone in univariate analyses (P < 0.0001). In adjusted analyses, patients randomized to rate control spent a mean of 42.2 ± 1.9% of their time in AF compared to 15.0 ± 1.8% for patients receiving amiodarone (P < 0.0001). In amiodarone treated patients, LVEF was not significantly associated with proportion of time spent in AF. The mean proportion of time spent in AF was 14.3 ± 2.4%, 16.2 ± 2.2%, and 14.9 ± 2.8% in patients with an LVEF <30%, 30–49%, and ≥50%, respectively (P = 0.6094). Figure 3 depicts AF burden in patients on rate control and treated with amiodarone according to their LVEF.
(Enlarge Image)
Figure 3.
Adjusted atrial fibrillation burden according to treatment strategy and left ventricular systolic function. Shown is the mean proportion of time spent in atrial fibrillation stratified by treatment modality (i.e., rate control or rhythm control with amiodarone) and left ventricular ejection fraction (LVEF). A significantly higher burden of atrial fibrillation was observed with rate control compared to amiodarone therapy (P < 0.0001), with no differences according to LVEF.
During follow-up, 1963 (59.4%) of 3,307 patients required at least one hospitalization, 1,401 (42.6%) of whom were hospitalized for a cardiovascular reason. In multivariate analyses, amiodarone was not associated with all-cause or cardiovascular hospitalizations in all patients (P = 0.2473 and P = 0.4347, respectively), nor in subgroups according to presence or absence of severe left ventricular dysfunction (hazard ratios for interactions 0.86, 95% CI [0.63–1.19], P = 0.3641 and 0.72, 95% CI [0.50–1.02], P = 0.0643, respectively). Table 3 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9 summarizes the factors associated with all-cause and cardiovascular hospitalizations according to LVEF in multivariate analyses. No significant interaction was noted between the study of enrolment (i.e., AFFIRM versus AF-CHF) and amiodarone in the overall population and in subgroups with or without severe left ventricular systolic dysfunction.
Over the course of the study, 729 (22.0%) deaths occurred, 68.3% of which were classified as cardiovascular (N = 498; 15.1%). A total of 450 (20.5%) deaths, 298 (13.5%) cardiovascular, occurred in the 2,200 patients randomized to rate control. The remaining 279 (25.2%) deaths, 200 (18.1%) cardiovascular, were in the 1,107 patients on amiodarone. In multivariate analyses, amiodarone was not predictive of overall (P = 0.8749) or cardiovascular (P = 0.8997) mortality in all patients, nor in those with or without severe left ventricular systolic dysfunction (hazard ratios for interactions 1.03, 95% CI [0.68–1.55], P = 0.8874 and 1.02, 95% CI [0.63–1.65], P = 0.9328, respectively). Factors associated with all-cause and cardiovascular mortality according to LVEF are summarized in Table 4 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9. All interaction terms tested with amiodarone, including study of enrollment, presence of ischemic heart disease, and NYHA functional class, were not significantly associated with all-cause or cardiovascular mortality in all patients and in the subgroups with and without severe left ventricular systolic dysfunction.
Results
Study Population
A total of 3,307 patients enrolled in AFFIRM and AF-CHF trials met inclusion criteria (Fig. 1). Overall, 2,200 patients were randomized to rate control and 1,107 to rhythm control with amiodarone as the first antiarrhythmic agent. Baseline characteristics are summarized in Table 1. The average age was 68.0 ± 0.2 years and 1,027 (31.1%) patients were female. Coronary artery disease was prevalent in 1,441 (43.6%), diabetes in 681 (20.6%), and hypertension in 2,068 (62.5%) patients.
(Enlarge Image)
Figure 1.
Selection of study population from AFFIRM and AF-CHF trials. LVEF = left ventricular ejection fraction.
For patients with an LVEF <30% (N = 842), no difference in baseline characteristics were observed according to whether they were randomized to rate control (N = 476) or rhythm control with amiodarone (N = 366). In contrast, for the subgroup with an LVEF ≥30% (N = 2,483), patients treated with amiodarone (N = 741) were more likely than those randomized to rate control (N = 1,724) to be male, have a higher NYHA functional class, coronary artery disease, and receive β-blockers, lipid lowering drugs, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers.
Recurrent AF on Amiodarone Therapy
Over the course of the study, 274 of 1,107 (24.7%) patients initially assigned to amiodarone discontinued the drug. Reasons for discontinuation are listed in Table 2 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9. Recurrence of AF was assessed in eligible patients randomized to rhythm control who received amiodarone as the first antiarrhythmic agent (N = 713; Fig. 1). Baseline characteristics of this patient subgroup are provided in Table 2 according to whether the LVEF was <30% (N = 237), 30–49% (N = 244), or ≥50% (N = 232). Patients with lower LVEFs were younger, more likely to be male, and had a higher prevalence of underlying heart disease and associated comorbidities.
Over an average follow-up of 40.0 ± 0.3 months, 272 (38.1%) patients experienced recurrent AF. Corresponding recurrence-free survival rates were 84%, 72%, and 45% at 1, 2, and 5 years, respectively. Kaplan–Meier AF-free survival curves are shown in Figure 2. No differences in rates of recurrent AF were observed according to LVEF in univariate (P = 0.6306) or multivariate (P = 0.8754) analyses. Univariate and multivariate predictors of recurrent AF are summarized in Table 3, along with nonsignificant selected variables of interest. The only variables independently associated with recurrent AF were diabetes (hazard ratio 1.60, 95% confidence interval CI: [1.15–2.17], P = 0.0044 and larger left atrial size (hazard ratio 1.03 per 1 mm increase, 95% CI [1.01–1.05], P = 0.0096.
(Enlarge Image)
Figure 2.
Freedom from recurrent atrial fibrillation according to left ventricular systolic function. Shown are Kaplan–Meier curves depicting freedom from recurrent atrial fibrillation in the 713 patients randomized to rhythm control, in sinus rhythm at time 0, and receiving amiodarone as the first-line antiarrhythmic agent. No statistically significant difference in freedom from recurrent atrial fibrillation was observed according to left ventricular ejection fraction (LVEF) stratified as ≥50%, 30–49%, and <30%.
Patients who were excluded from the analysis of AF recurrence by virtue of failing to achieve normal sinus rhythm at time zero (N = 345) had a similar all-cause mortality rate to those in sinus rhythm (adjusted P = 0.6875) but a higher all-cause hospitalization rate (adjusted P < 0.0001). The lack of association between LVEF and AF recurrence was corroborated by a sensitivity analysis that included the 345 patients in AF at the 4-month visit by considering AF to have recurred at day 1 (P = 0.2075). Similarly, forcing the variable "rhythm at the time of echocardiography" into multivariate models had no appreciable effect on outcomes.
AF Burden
The proportion of time spent in AF was significantly higher in patients randomized to rate control therapy compared to rhythm control with amiodarone in univariate analyses (P < 0.0001). In adjusted analyses, patients randomized to rate control spent a mean of 42.2 ± 1.9% of their time in AF compared to 15.0 ± 1.8% for patients receiving amiodarone (P < 0.0001). In amiodarone treated patients, LVEF was not significantly associated with proportion of time spent in AF. The mean proportion of time spent in AF was 14.3 ± 2.4%, 16.2 ± 2.2%, and 14.9 ± 2.8% in patients with an LVEF <30%, 30–49%, and ≥50%, respectively (P = 0.6094). Figure 3 depicts AF burden in patients on rate control and treated with amiodarone according to their LVEF.
(Enlarge Image)
Figure 3.
Adjusted atrial fibrillation burden according to treatment strategy and left ventricular systolic function. Shown is the mean proportion of time spent in atrial fibrillation stratified by treatment modality (i.e., rate control or rhythm control with amiodarone) and left ventricular ejection fraction (LVEF). A significantly higher burden of atrial fibrillation was observed with rate control compared to amiodarone therapy (P < 0.0001), with no differences according to LVEF.
Hospitalizations
During follow-up, 1963 (59.4%) of 3,307 patients required at least one hospitalization, 1,401 (42.6%) of whom were hospitalized for a cardiovascular reason. In multivariate analyses, amiodarone was not associated with all-cause or cardiovascular hospitalizations in all patients (P = 0.2473 and P = 0.4347, respectively), nor in subgroups according to presence or absence of severe left ventricular dysfunction (hazard ratios for interactions 0.86, 95% CI [0.63–1.19], P = 0.3641 and 0.72, 95% CI [0.50–1.02], P = 0.0643, respectively). Table 3 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9 summarizes the factors associated with all-cause and cardiovascular hospitalizations according to LVEF in multivariate analyses. No significant interaction was noted between the study of enrolment (i.e., AFFIRM versus AF-CHF) and amiodarone in the overall population and in subgroups with or without severe left ventricular systolic dysfunction.
Mortality
Over the course of the study, 729 (22.0%) deaths occurred, 68.3% of which were classified as cardiovascular (N = 498; 15.1%). A total of 450 (20.5%) deaths, 298 (13.5%) cardiovascular, occurred in the 2,200 patients randomized to rate control. The remaining 279 (25.2%) deaths, 200 (18.1%) cardiovascular, were in the 1,107 patients on amiodarone. In multivariate analyses, amiodarone was not predictive of overall (P = 0.8749) or cardiovascular (P = 0.8997) mortality in all patients, nor in those with or without severe left ventricular systolic dysfunction (hazard ratios for interactions 1.03, 95% CI [0.68–1.55], P = 0.8874 and 1.02, 95% CI [0.63–1.65], P = 0.9328, respectively). Factors associated with all-cause and cardiovascular mortality according to LVEF are summarized in Table 4 of the online data supplement http://onlinelibrary.wiley.com/store/10.1111/jce.12535/asset/supinfo/jce12535-sup-0001-supinfo.docx?v=1&s=8d071af8eb09733f39c5cc232f16bcc7b4f755e9. All interaction terms tested with amiodarone, including study of enrollment, presence of ischemic heart disease, and NYHA functional class, were not significantly associated with all-cause or cardiovascular mortality in all patients and in the subgroups with and without severe left ventricular systolic dysfunction.
Source...