Author + information
- Received June 3, 1998
- Revision received November 5, 1998
- Accepted March 15, 1999
- Published online August 1, 1999.
- Thomas Guarnieri, MD, FACCa,* (, )
- Shellee Nolan, MD, FACCa,
- Sidney O Gottlieb, MD, FACCa,
- Audrey Dudek, RN, MSNa and
- David R Lowry, DrPHa
- ↵*Reprint requests and correspondence: Dr. Thomas Guarnieri, Midatlantic Cardiovascular Associates, 6569 North Charles Street, Baltimore, Maryland 21204
This study was designed to test whether intravenous (IV) amiodarone would prevent atrial fibrillation and decrease hospital stay after open heart surgery.
Atrial fibrillation commonly occurs after open heart procedures and is thought to be a significant determinant for prolongation of hospitalization. Oral amiodarone given preoperatively appears to reduce the incidence of atrial fibrillation. This study was designed to test whether the more rapid-acting IV formulation of amiodarone given postoperatively would reduce the incidence of atrial fibrillation.
Three hundred patients undergoing standard open heart surgery were randomized in a double-blind fashion to IV amiodarone (1 g/day for 2 days) versus placebo immediately after open heart surgery. The primary end points of the trial were incidence of atrial fibrillation and length of hospital stay. Baseline clinical variables and mortality and morbidity data were collected.
Atrial fibrillation occurred in 67/142 (47%) patients on placebo versus 56/158 (35%) on amiodarone (p = 0.01). Length of hospital stay for the placebo group was 8.2 ± 6.2 days, and 7.6 ± 5.9 days for the amiodarone group (p = 0.34). No differences were noted in baseline variables, morbidity or mortality.
Low-dose IV amiodarone was safe and effective in reducing the incidence of atrial fibrillation after heart surgery, but did not significantly alter length of hospital stay.
Atrial fibrillation frequently occurs after open heart surgery, with estimates ranging to 50% (1–7). Matthew et al. (2)recently reported the results from one of the largest studies on postoperative atrial fibrillation. The MultiCenter Study of Perioperative Ischemia demonstrated a frequency of atrial fibrillation following coronary bypass grafting of 27%. Multiple risk factors were identified including age, heart failure and some surgical practices. Moreover, hospital
length of stay was significantly prolonged in those patients who experienced atrial fibrillation. Other investigators have reported similar frequencies, risk factors and prolongation of hospitalizations (8–15).
A number of strategies involving pharmacologic prophylaxis have been used to prevent atrial fibrillation after bypass surgery (16–27). A recent report suggests that large oral doses of amiodarone may be effective for reducing both the incidence of postcoronary bypass atrial fibrillation and length of hospital stay (28). Daoud et al. (28)utilized several days of preoperative oral amiodarone to load the patients with the drug. Alternatively, low-dose intravenous (IV) amiodarone has been demonstrated to treat refractory ventricular tachycardia in hours as opposed to days for the oral preparation (29,30). Because the IV formulation appears to act so rapidly, we initiated the ARCH study (Amiodarone Reduction in Coronary Heart study) to examine whether IV postoperative amiodarone will prevent atrial fibrillation and reduce length of hospital stay.
The ARCH study was designed to test the hypotheses that low-dose IV amiodarone, given postoperatively, would reduce the frequency of atrial fibrillation and the length of hospital stay. The ARCH study was conducted as a randomized double-blind trial of IV amiodarone versus placebo at a single surgical center in 300 patients with no history of atrial fibrillation undergoing standard open heart procedures.
From September 9, 1996, to August 12, 1997, patients undergoing standard open heart surgery with no evidence of preoperative atrial fibrillation were screened for participation in this study, which was reviewed and approved by the St. Joseph Medical Center Institutional Review Board. All patients signed informed consent. Three hundred patients were enrolled. All patients underwent standard open heart procedures via midline sternotomy. We excluded patients undergoing other methods of cardiac surgery, including lateral thoracotomy, “mid-cab,” or various thoracoscopy approaches. Patients were excluded if they had any history of atrial fibrillation or were taking class I or class III antiarrhythmic drugs. Patients were included if they were taking digoxin, beta-blockers or calcium channel blockers for reasons other than any cardiac arrhythmia. As intended, drug was not infused in randomized patients if they arrived in the surgical intensive care unit in atrial fibrillation, had an aortic counterpulsation balloon in place or the first two cardiac indexes (determined by thermodilution technique at least 4 h apart) were less than 2.0 liter/min/m2. These randomized but not infused patients were included in the analysis as intention to treat.
Randomization and drug administration
Patients were randomly assigned in a double-blind fashion to receive either IV amiodarone or placebo infused through a central venous catheter. A 1:1 treatment allocation scheme was used in this study. The amiodarone was infused at a rate of 1 g over 24 h, given for a total of 48 h (2 g total). No loading dose was given. The drug infusion was started within 3 h of entering the surgical intensive care unit. The only hemodynamic parameter for terminating infusion was systolic blood pressure consistently <80 mm Hg. Because all patients received temporary atrial and ventricular pacing wires, bradycardia was not a parameter for study termination.
Study end points
The primary end points of this study were the development of postoperative atrial fibrillation, detected via continuous ambulatory monitoring, and hospital length of stay. Atrial fibrillation was defined as any episode of atrial fibrillation requiring any treatment. Treatment for atrial fibrillation was not controlled. Patients were all followed up to one month after discharge. At the one month visit, rhythm analysis was performed. The investigators did not have access to the data until the trial was complete and analysis had been concluded. The data reported are based on 100% level of completeness for length of hospital stay and incidence of atrial fibrillation.
Data were analyzed using the intention-to-treat principle. Continuous data are summarized as means ± SD. Discrete variables were compared between treatment groups by the chi-square test, and continuous variables were compared by means of the Student ttest. Kaplan-Meier analysis with the log-rank test was used to compare the probability of postoperative atrial fibrillation by treatment group.
A total of 300 patients were randomized in this study. Over 28 baseline characteristics were prospectively collected. There were no differences in the treatment groups. Representative baseline characteristics, displayed in Table 1, reflect the similarity in the two groups. Specifically, age, indexes of ventricular function and incidence of co-morbid illness were identical in the two groups. It should be noted that about one-third of the population was older than 70 years, and a significant proportion had left ventricular dysfunction. Approximately 50% of both populations were taking some formulation of beta-adrenergic blocking agent before surgery (Table 1). By the time of discharge, 18.1% (25 of 142) of the control group and 11.4% (17 of 158) of the amiodarone-treated group had beta-blockers discontinued (p = 0.18). Operative characteristics are displayed in Table 2. There were no differences in the operative interventions or techniques between the two treatment groups. One patient in the placebo group (listed as “other”) had an atrial septal defect repair. In this population, there were no differences in location of grafts or vessels bypassed.
Study drug was not infused in 14 (5%) of the 300 randomized patients: 6 patients in the placebo group and 8 in the amiodarone group received no infusion because of immediate postoperative hemodynamic instability. The study drug was terminated prematurely in seven (4.4%) of the patients assigned to amiodarone and in two (1.4%) patients assigned to placebo (p = 0.12) (Table 3). All patients were included in the analysis as intention to treat. Because all patients had temporary epicardial atrial and ventricular pacing wires placed, bradycardia was not an indication for infusion termination. Nonetheless, two patients had failure of capture of the atrial leads during mild bradycardia, for which the infusion was terminated.
Atrial fibrillation developed in 67 of 142 patients assigned to placebo (47% incidence) and in 56 of 158 (35%) patients assigned to amiodarone (p = 0.01), representing a 26% relative risk reduction. The percentages of patients free of atrial fibrillation during hospitalization by treatment group are shown in Figure 1, and confirm a significant reduction in atrial fibrillation in the amiodarone group. Analysis of the Kaplan-Meier curve demonstrates that the drug effect appears to occur in the first few days of treatment. The first episode of atrial fibrillation occurred on postoperative day 2.9 ± 1.2 in the group receiving amiodarone and on day 2.3 ± 1.3 in the placebo group (p < 0.01). The mean ventricular response rate during at least 30 s of the first episode of atrial fibrillation was 135.5 ± 31.3 beats/min in the placebo group and 128.8 ± 28.7 beats/min in the amiodarone-treated group (p = 0.26). Average hospital charges for the placebo group were $18,889; for the amiodarone group (p = 0.18), hospital charges averaged $16,871.
The length of hospital stay did not statistically differ between treatment groups: 7.6 ± 5.9 days for the amiodarone group versus 8.2 ± 6.2 days for the placebo group (p = 0.34). As in other studies, atrial fibrillation was associated with a prolonged length of hospital stay regardless of treatment assignment. Patients who experienced atrial fibrillation had a length of stay of 9.0 ± 5.1 days; the length of hospital stay in those patients free of atrial fibrillation was 7.1 ± 6.6 days (p < 0.01).
Nonfatal postoperative complications occurred in four (2.5%) patients who received amiodarone and in seven (4.9%) patients receiving placebo (p = 0.27). No deaths (0%) occurred in the amiodarone group, whereas two deaths (1.4%) occurred among patients in the placebo group (p = 0.13) (Table 4).
At the four- to six-week follow-up, 236 patients were in sinus rhythm, 5 were in atrial fibrillation and 59 had no electrocardiographic documentation of rhythm.
Atrial fibrillation after open heart surgery
Atrial fibrillation remains one of the most frequent complications of coronary bypass surgery. Matthew et al. (2)reported the outcomes of over 2,000 patients examined in the Multicenter Study of Perioperative Ischemia. In that multicenter database, the incidence of atrial fibrillation was 27%. Aranki et al. (1)reported the results on over 500 patients from a single center. In that series the incidence was 33%. Both of these large series and several smaller ones suggest the incidence of atrial fibrillation may range from 25% to 50% (1,2,28).
Both Matthew and Aranki also provided strong evidence that atrial fibrillation significantly prolonged hospitalization and increased resource utilization (1,2). In the Perioperative Ischemia study, both prolonged hospitalization (two days) and extra time in the intensive care unit (13 h) were noted. Aranki (1)demonstrated that atrial fibrillation was associated with an extra 4.9 days of hospitalization with estimated additional charges of $10,055.
Amiodarone and atrial fibrillation after open heart surgery
The ARCH study is the largest randomized trial of IV amiodarone given as prophylaxis for postoperative atrial fibrillation. The primary finding of this study is that low dose IV amiodarone decreases the incidence of atrial fibrillation after open heart surgery by 26%. The protocol was designed specifically to analyze the effects of the IV formulation of amiodarone given only in the immediate postoperative period. Unlike the oral compound, which has an onset of action of at least several days, IV amiodarone appears to act within a few hours as judged by the treatment of refractory ventricular tachycardia (29,30). We speculated that the rapid onset of the IV formulation of amiodarone could provide enough drug to prevent atrial fibrillation in the days immediately postbypass when atrial fibrillation was manifest. Because many of the patients in this series were identified as urgent candidates for bypass and were not seen in the outpatient clinic before surgery, the concept of administering drug only in the immediate postoperative period was both attractive and practical.
Amiodarone has been used before to prevent atrial fibrillation in small series, often combining the oral and IV formulation. Butler et al. (31)administered 15 mg/kg of amiodarone immediately after releasing the aortic cross clamp followed by 600 mg/day for three days. Eight percent of the amiodarone-treated group had “clinically significant” supraventricular arrhythmias as compared with 20% in the placebo control group. Hohnloser and colleagues (32)administered a loading IV bolus of 300 mg of amiodarone postoperatively followed by 1200 mg/day for two days and 900 mg/day for two days. Eight of 38 control patients (21%) experienced atrial fibrillation while only 2 of 39 (5%) patients treated with amiodarone had atrial fibrillation. The ARCH study demonstrated a similar risk reduction. More recently, Daoud (28)identified patients preoperatively and gave 600 mg orally per day for seven days before surgery followed by 200 mg/day until discharge. In the control group, 32 of 64 (50%) experienced atrial fibrillation, whereas 16 of 64 (25%) treated with amiodarone had atrial fibrillation. The degree of benefit appears to be greatest in the Daoud study, possibly because of the large preoperative dose, 4.8 g. All of these studies, including the ARCH study, have shown significant reduction in the incidence of atrial fibrillation after open heart surgery with prophylactic amiodarone.
Atrial fibrillation and length of hospital stay
Several studies have examined whether atrial fibrillation is an important risk factor for prolonging hospitalization. In the large Multicenter Study of Perioperative Ischemia, 2,417 patients were studied (2). Patients with atrial fibrillation had a 2.0-day longer hospitalization than did those without atrial fibrillation. Similarly, Aranki (1)reported a single-center series of 570 patients. The length of stay for patients with atrial fibrillation was 15 days, compared with 9 days in those without atrial fibrillation (1). Both studies strongly supported the hypothesis that atrial fibrillation was an important independent variable in length of hospital stay. Daoud was the first to demonstrate that amiodarone treatment resulted in a shortening of hospitalization: 7.9 days for the control and 6.5 days for the treated group (28). Daoud and co-workers proposed that the decrease in hospitalization reduced costs from approximately $26,000 in the control group to $18,000 in the treated group. The ARCH study showed a nonsignificant but similar trend in reducing hospitalization, 8.2 days for controls and 7.5 days for the treatment group.
Complications of therapy
None of the prospective trials of prophylactic amiodarone has shown any significant mortality or morbidity from the drug (28,31,32). Previous trials had relatively small numbers and may not have been powered to detect what is probably a small drug risk. In the ARCH study, which is the largest to date, we found no evidence of morbidity or mortality secondary to therapy. Certainly, the short-term exposure to rather modest total doses of amiodarone may be the most important factor in reducing drug risk.
The ARCH study was specifically designed to test the effect of low doses of IV amiodarone in preventing atrial fibrillation when given postoperatively. It should be noted that atrial fibrillation tended to occur at day 3 in the treated group, after the termination of the infusion. Certainly, a strategy could have been used to give higher total doses, either by increases in concentration or length of infusion, or addition of oral compound. Given the results of the Daoud report (28), it is possible that higher total doses might have a more pronounced effect. The ARCH study did not demonstrate a significant decrease in hospitalization, one of the primary end points. The study was powered to detect a 20% reduction in atrial fibrillation, but it may not have been sufficiently powered to detect differences in hospitalization because of the large variances involved.
Intravenous amiodarone given in the immediate postoperative period is safe and effective in reducing the incidence of atrial fibrillation. The degree of risk reduction was not sufficient to produce a significant decrease in hospitalization.
- Received June 3, 1998.
- Revision received November 5, 1998.
- Accepted March 15, 1999.
- American College of Cardiology
- Aranki S.F,
- Shaw D.P,
- Adams D.H,
- et al.
- Caretta Q, Mercanti CA, De Nardo D, et al. Ventricular conduction defects and atrial fibrillation after coronary artery bypass grafting: multivariate analysis of preoperative, intraoperative and postoperative variables. Eur Heart J 1991;12:1107–1.
- Ormerod O.J.M,
- McGregor C.G.A,
- Stone D.L,
- Wisbey C,
- Petch M.C
- Johnson L.W,
- Dickstein R.A,
- Fruehan C.T
- Lamb R.K,
- Prabhnkar G,
- Thorpe J.A.C,
- Smith S,
- Norton R,
- Dyde J.A
- Scheinman M.M,
- Levine J.H,
- Cannom D.S,
- et al.
- Levine J.H,
- Massumi A,
- Scheinmann M.M,
- et al.
- Butler J,
- Harriss D.R,
- Sinclair M,
- Westaby S