Author + information
- Received August 16, 1999
- Revision received December 16, 1999
- Accepted February 2, 2000
- Published online May 1, 2000.
- Michael D Greenberg, MDa,
- Nevin M Katz, MDa,
- Stephen Iuliano, MDa,
- Barbara J Tempesta, RNa and
- Allen J Solomon, MD, FACCa,* ()
- ↵*Reprint requests and correspondence: Dr. Allen J. Solomon, Division of Cardiology, Georgetown University Medical Center—Room M4222, 3800 Reservoir Rd, NW, Washington, DC 20007
The purpose of this study was to determine the efficacy of atrial pacing in the prevention of atrial fibrillation following cardiovascular surgery.
Although pharmacologic therapy has been used to help prevent postoperative atrial fibrillation, it suffers from limited efficacy and adverse effects. In the nonoperative setting, novel pacing strategies have been shown to reduce recurrences of atrial fibrillation and prolong arrhythmia-free periods in patients with paroxysmal atrial arrhythmias.
A total of 154 patients (115 men; mean age, 65 ± 10 years; ejection fraction, 53 ± 10%) undergoing cardiac surgery (coronary artery bypass surgery, 88.3%; aortic valve replacement, 4.5%; coronary bypass + aortic valve replacement, 7.1%) had right and left atrial epicardial pacing electrodes placed at the time of surgery. Patients were randomized to either no pacing, right atrial (RAP), left atrial (LAP) or biatrial pacing (BAP) for 72 h after surgery. Beta-adrenergic blocking agents were administered concurrently to all patients following surgery.
There was a reduction in the incidence of postoperative atrial fibrillation from 37.5% in patients receiving no postoperative pacing to 17% (p < 0.005) in patients assigned to one of the three pacing strategies. The length of hospital stay was reduced by 22% from 7.8 ± 3.7 days to 6.1 ± 2.3 days (p = 0.003) in patients assigned to postoperative atrial pacing. The incidence of atrial fibrillation was lower in each of the paced groups (RAP, 8%; LAP, 20%; BAP, 26%) compared with patients who did not receive postoperative pacing (37.5%).
Postoperative atrial pacing, in conjunction with beta-blockade, significantly reduced both the incidence of atrial fibrillation and the length of hospital stay following cardiovascular surgery. Additional studies are needed to determine the most effective anatomic pacing site.
Atrial fibrillation (AF) is a common complication in the setting of cardiovascular surgery. Of the estimated 300,000 patients undergoing coronary artery bypass grafting (CABG) surgery annually, between 20% and 40% will have development of AF in the week following surgery (1,2). The occurrence of postoperative AF has significant implications both in terms of morbidity and with respect to utilization of
hospital resources. In addition, there is a higher incidence of postoperative congestive heart failure, renal insufficiency, infection and neurologic injury in patients with postoperative AF (3). The prolonged length of hospital stay and increased hospital charges that accompany postoperative AF are a major contributing factor to the health-care costs associated with coronary artery bypass surgery.
Beta-adrenergic blocking agents have consistently been shown to reduce the incidence of postoperative AF (1,4–9). Two meta-analyses found beta-blockers to be highly effective in preventing the occurrence of postoperative AF (1,9). In fact, beta-blockers have had a beneficial effect on postoperative AF in nearly every study performed. More recently, several studies have demonstrated the potential efficacy of amiodarone and sotalol as prophylactic agents for the prevention of postoperative AF (10–12).
A growing body of data, in the nonoperative setting, have suggested that innovative pacing therapies such as dual site atrial pacing may decrease the incidence of paroxysmal AF (13,14). As a result, we looked at the feasibility of atrial pacing following coronary artery bypass surgery. Preliminary reports indicate that atrial-based pacing may be performed safely in the postoperative setting, but data regarding the effectiveness of postoperative pacing in reducing the incidence of postoperative AF are scant. In a recent study by Gerstenfeld and colleagues (15), no significant reduction in postoperative AF was found with either right atrial (RAP) or biatrial pacing (BAP); however, a trend toward a reduction in AF was observed in patients who received beta-blockers in conjunction with postoperative pacing (15).
The present study was designed to assess the efficacy of prophylactic atrial-based pacing, in conjunction with beta-blockers, in the prevention of postoperative AF. In addition, we sought to determine if there was a selective benefit to RAP, left (LAP), or BAP in the prevention of AF following cardiac surgery.
Patients 18 years of age and older who were scheduled to undergo either CABG or aortic valve surgery were screened for participation in the study. Informed consent was obtained from all patients in accordance with a protocol that had been approved by the Institutional Review Board at Georgetown University prior to surgery. Patients were randomly assigned to one of four strategies: no postoperative pacing (NAP), RAP, LAP, or BAP. Patients were excluded from the study if there was a documented history of AF in the 24 h preceding surgery, if they were receiving an antiarrhythmic agent or if they were not in sinus rhythm at the time of surgery.
Surgical procedure and postoperative pacing protocol
Patients underwent cardiovascular surgery utilizing standard surgical techniques. After completion of the surgical procedure, but before discontinuation of cardiopulmonary bypass, the lateral wall of the right atrium and the right lateral aspect of the left atrium were exposed. Two epicardial pacing electrodes were placed into the left atrial myocardium just inferior to the origin of the right superior pulmonary vein. The RAP electrodes were then placed in a similar manner on the posterior aspect of the right atrium where the atrial wall is thicker. Finally, a set of electrodes was attached to the right anterior ventricular wall. The pacing electrodes were held in place with absorbable clips. The atrial electrodes were brought out from the mediastinum to the right of the median sternotomy incision. The electrodes were secured to the skin at their exit sites using skin sutures.
On return from the operating room, all patients had their epicardial pacing electrodes connected to an external temporary dual-chamber pulse generator (Medtronic model 5345; Minneapolis, Minnesota). The pair of RAP electrodes were connected to the atrial ports of the pulse generator, while the pair of LAP electrodes were connected to the ventricular ports. When RAP was desired, the pacemaker was programmed in the AAI mode, whereas it was programmed in the VVI mode for LAP. For BAP, the pacemaker was programmed in the DDD mode with the atrioventricular delay set to its lowest value (i.e., 6 ms). The pacing rate was then programmed based on the native heart rate (Table 1). Patients in the NAP group were not paced unless clinically indicated (pulse <50 beats/min or systolic blood pressure <100 mm Hg). RAP and LAP thresholds as well as sensitivity were determined every 8 h, and outputs were programmed accordingly to give a twofold safety margin.
In accordance with the randomization protocol, pacing was started as soon as possible on returning from the operating room and was continued for 72 h. The paced rate was adjusted whenever the native heart rate equaled or exceeded the paced rate (Table 1). Patients underwent continuous electrocardiographic monitoring for at least 96 h following surgery utilizing a centralized cardiac telemetry system (Marquette; Milwaukee, Wisconsin) equipped with arrhythmia and rate detection alarm triggers. After 96 h, patients were monitored through a combination of telemetry and daily electrocardiograms. Two “blinded” cardiologists reviewed all recorded electrocardiographic data daily. Beta-blockers were administered postoperatively to all patients unless contraindicated by bradycardia (heart rate <60 beats/min), heart block, hypotension (systolic blood pressure <90 mm Hg) or significant bronchospasm. Patients received a 5-mg test dose of propranolol orally immediately following extubation and the dose was titrated to a target dose of 80 mg daily (20 mg orally four times daily).
The primary end point of the study was the occurrence of AF lasting >1 h or resulting in hemodynamic compromise necessitating electrical or chemical cardioversion. The secondary end point was the length of hospital stay that was defined as the length of time from surgery until hospital discharge.
A sample size of 150 patients was required to detect a reduction in postoperative AF from 40% in nonpaced patients to 15% in patients receiving a pacing strategy (using a significance level of 5% and a power of 80%). All data were initially analyzed on an intention-to-treat basis. An on-treatment analysis was then performed on patients who had received continuous pacing for 72 h. Statistical analyses were by two-tailed Student’s t test for continuous data and chi-square for categorical data. Comparisons between groups were performed with analysis of variance with adjustment for multiple comparisons. Statistics were calculated using software (SPSS version 8.0; SPSS Inc; Chicago, Illinois). Statistical significance was considered present at a p value of <0.05.
A total of 154 patients met the inclusion criteria and were included in the study. There were 115 men (73%) with a mean age of 65 ± 10 years. The mean ejection fraction determined preoperatively by cardiac catheterization was 53 ± 10%. One hundred thirty-six patients underwent CABG surgery alone, 7 patients had isolated aortic valve replacement and the remaining 11 patients underwent combined CABG and aortic valve surgery. For the group as a whole, intraoperative aortic cross clamp time was 75 ± 25 min while bypass time was 115 ± 31 min. There were no significant differences in either the baseline characteristics (Table 2)or operative data (Table 3) for patients in the NAP, RAP, LAP and BAP groups.
Postoperative beta-blockers and feasibility of pacing
Postoperative beta-blocker therapy was started in all patients without contraindications. There was no significant difference between the groups with respect to the percentage of patients receiving beta-blockers, the time until receipt of the first dose of beta-blocker following surgery or the mean dose of beta-blockers received during the postoperative period (Table 4).
There were significant differences in the ability to maintain continuous pacing among the different groups. Right atrial pacing was maintained for the entire 72-h period without difficulty in most patients (90%); however, both LAP and BAP were more problematic. Inability to maintain pacing either as a result of diaphragmatic stimulation or high thresholds occurred in 23% of patients assigned to LAP and 33% of those assigned to BAP. In addition, 19% of patients randomized to the NAP group required some form of postoperative pacing due to bradycardia, heart block or hypotension. There were no complications associated with the placement, maintenance or removal of the atrial pacing electrodes.
Efficacy of postoperative pacing
As a group, patients who were assigned to one of the three pacing strategies had a significantly lower incidence of postoperative AF compared with patients who were randomized to an NAP strategy (17% vs. 37.5%; p < 0.005). The mean time from surgery until the onset of postoperative AF was 54 ± 44 h and it did not differ significantly between groups.
The proportion of patients developing postoperative AF was lower in each of the paced groups when compared with patients assigned to NAP (Fig. 1). Atrial fibrillation occurred in 8% of patients assigned to RAP, 20% with LAP and 26% with BAP. This was in contrast to the 37.5% incidence of postoperative AF present in patients assigned to the NAP strategy.
Comparisons among each of the three pacing strategies demonstrated that patients assigned to RAP had a 79% reduction in the incidence of postoperative AF (RAP, 8%; NAP, 37.5%; p = 0.002). Although the incidence of postoperative AF was lower in both the LAP group (NAP vs. LAP; p = 0.14) and the BAP group (NAP vs. BAP; p = 0.40), this did not reach statistical significance.
When the data were analyzed using an on-treatment analysis, the results were similar to those obtained by the intention-to-treat analysis. The incidence of postoperative AF among patients who actually received 72 h of pacing therapy was 13% (RAP, 11%; LAP, 15%; BAP, 24%) compared with 33% (p < 0.005) in those in the NAP group.
Length of hospital stay
Overall, patients assigned to a pacing strategy following surgery had a 22% reduction in their length of hospital stay compared with patients who were in the NAP group (6.1 ± 2.3 days vs. 7.8 ± 3.7 days; p = 0.003). Compared with a mean length of stay of 7.8 ± 3.7 days for patients assigned to NAP, the mean length of stay was 5.9 ± 2.1 days for BAP, 6.8 ± 2.9 days for LAP and 5.6 ± 1.4 days for RAP. When analyzed by individual groups, the difference between NAP and RAP was significant (p = 0.01); there was a trend toward a decrease in hospital stay between NAP and BAP (p = 0.08) and no significant difference between NAP and LAP (Fig. 2).
The results of the present study have demonstrated that postoperative atrial pacing in conjunction with beta-blockers significantly reduced the incidence of AF following cardiac surgery. This strategy also significantly shortened the length of hospital stay.
Conventional treatment strategies
Atrial fibrillation is the most common postoperative complication following cardiac surgery (2). Although it usually does not result in long-term sequelae, it does add significantly to morbidity, utilization of hospital resources and the length of hospitalization (2,3). Treatment strategies have, until recently, centered on the use of pharmacologic therapy. Beta-blockers and more recently several of the class III antiarrhythmic agents have been shown to be effective in reducing the incidence of postoperative AF (4–12). However, pharmacotherapy has been limited by adverse effects and a significant occurrence of AF even in patients receiving optimal treatment.
Previous studies of atrial pacing for the prevention of AF
The limited efficacy of conventional agents has led to searches for nonpharmacologic modalities for the prevention of postoperative AF. Coumel and colleagues (16,17) were among the first to report the potential of pacing to prevent AF almost a decade ago. They described the use of single-site atrial overdrive pacing to prevent AF or flutter in a select group of patients with vagally mediated AF or flutter (16,17). Further studies by Murgatroyd et al. (18) utilized a unique pacing algorithm for the suppression of atrial premature depolarizations. This technique resulted in a significant reduction in episodes of AF. The mechanism by which atrial overdrive pacing reduces the occurrence of AF is unclear, although suppression of atrial premature depolarizations and a reduction in the dispersion of refractoriness have all been proposed (19,20). The use of BAP was shown by Daubert and associates (13) to be effective in the prevention of atrial arrhythmias in patients with interatrial conduction block. By simultaneous stimulation from the right atrial appendage and the coronary sinus, the use of BAP was believed to resynchronize the atrial electrical activity and reduce intraatrial and interatrial asynchrony. Subsequent work by Saksena and colleagues (14) demonstrated the effectiveness of dual-site RAP in prolonging arrhythmia-free intervals in a group of patients with drug-refractory paroxysmal AF.
Despite the growing body of data demonstrating a role for atrial pacing in the prevention of AF in the nonoperative setting, its feasibility and efficacy following surgery remain less clear. Recently, Gerstenfeld et al. (15) reported no benefit with atrial pacing for the prevention of postoperative AF. They did, however, note that for the subset of patients who received both postoperative beta-blockers and atrial pacing, there was a reduction in the incidence of AF, suggesting that the combination of beta-blockade and atrial pacing might be of particular benefit.
The present study
Our study was designed to determine whether a treatment strategy of atrial-based pacing in conjunction with beta-blockers would reduce the incidence of AF following cardiac surgery. We demonstrated an overall reduction in postoperative AF from 37.5% in patients receiving no pacing to 17% in patients assigned to one of the postoperative pacing strategies. Of note, our 37.5% incidence of postoperative AF in patients receiving only beta-blockers, as well as the 8% incidence of AF in patients assigned to RAP, is similar to that recently reported by Gerstenfeld et al. (15). Although several investigators have suggested that BAP may be more efficacious than single-site pacing in reducing postoperative AF, in the present study, the greatest reduction in postoperative AF was seen in patients assigned to RAP. Although the present study clearly demonstrates the ability of postoperative pacing to reduce the occurrence of postoperative AF, inferences regarding a selective benefit of one form of pacing over another must be tempered in view of the significant differences among the groups with respect to the ability to initiate and maintain effective pacing. Because the ability to effectively pace was significantly greater in patients assigned to RAP, the lower incidence of AF in this group may be due either to a selective advantage of RAP or to the fact that RAP was more easily accomplished and maintained compared with LAP or BAP.
Although we have used the term BAP to imply simultaneous RAP and LAP, our actual technique involved pacing the right and left atria by using the atrial and ventricular ports of a temporary pacemaker with the atrioventricular delay set to its lowest value (6 ms). Therefore, there was actually a 6-ms delay between activation of the right and left atria. Although independent right and left atrial capture as well as changes in P-wave configuration were documented at the initiation of pacing and every 8 h, the possibility that biatrial capture was not maintained in all cases cannot be completely excluded.
Since continuous Holter monitoring was not performed on each patient, it is conceivable that some episodes of AF were undetected. Several precautions were taken to minimize this potential problem. Each patient underwent continuous electrocardiographic monitoring for at least 96 h following surgery. In conjunction with continuous monitoring, arrhythmia event–triggered alarms were preprogrammed for each patient. These sounded an audible alarm as well as triggering an ECG recording anytime there was an irregular heart rhythm or the heart rate exceeded a prespecified upper rate limit (i.e., 110 to 120 beats/min). All recorded events were retrieved for daily review. Routine rhythm ECG strips were collected by nursing personal every 1 to 6 h for all patients. Finally, every medical chart was reviewed separately by two physicians following patient discharge from the hospital. All enclosed ECG strips, nursing notes, vital signs and medication sheets were reviewed for any occurrences of AF.
While patients were randomized to the four different treatment groups, there was a substantial difference in the final number of patients in each group and while there were no differences in the baseline characteristics among the groups, the difference in group numbers must be viewed as a potential limitation.
We chose as our primary end point the occurrence of clinically significant AF. This was defined as either AF of 1 h duration or <1 h but associated with hypotension, ECG changes or the need for cardioversion. The possibility that we did not detect short, self-limiting episodes of AF cannot be dismissed, although to our knowledge, there are no data to suggest that such episodes are clinically significant.
Finally, while pacing from the right atria was easily initiated and maintained in the vast majority of patients, the same was not true for both LAP and BAP. In a significant proportion of cases, LAP and BAP could not be sustained because of high pacing thresholds or diaphragmatic stimulation. Unlike BAP in the nonoperative setting where LAP is achieved via the coronary sinus, in our study, the position of the left atrial epicardial electrodes probably resulted in stimulation of the phrenic nerve. This resulted in a significant group of patients who could not have effective pacing and hence the true clinical utility of this pacing strategy must be interpreted in this light.
Effective means for the prevention of postoperative AF have so far been limited to the use of beta-blockers and class III antiarrhythmic agents. The present study describes a safe and effective nonpharmacologic modality for the reduction of postoperative AF. The procedure described is easily applied in the clinical setting. The widespread utilization of postoperative atrial pacing could have important implications with respect to utilization of hospital resources, length of hospital stay and the complications associated with postoperative AF. Further studies are needed to better define the best anatomic site for postoperative pacing.
Prophylactic atrial pacing in conjunction with beta-blockers significantly reduced both the incidence of postoperative AF and the length of hospital stay following cardiac surgery.
We would like to acknowledge the efforts of the entire cardiothoracic team at Georgetown University. This study would not have been possible without the dedicated efforts of the nurses, anesthesiologists and surgical residents.
☆ This work was supported by a postdoctoral research fellowship award from the American Heart Association.
- atrial fibrillation
- biatrial pacing
- coronary artery bypass grafting
- left atrial pacing
- no pacing
- right atrial pacing
- Received August 16, 1999.
- Revision received December 16, 1999.
- Accepted February 2, 2000.
- American College of Cardiology
- Andrews T.C.,
- Relmold S.C.,
- Berlin J.A.,
- et al.
- Aranki S.F.,
- Shaw D.P.,
- Adams D.H.,
- et al.
- Ali I.M.,
- Sanalla A.A.,
- Clark V.
- Lamb R.K.,
- Prabhakar G.,
- Thorpe J.A.C.,
- et al.
- Guarnieri T.,
- Nolan S.,
- Gottlieb S.O.,
- et al.
- Gomes J.A.,
- Ip J.,
- Santoni-Rugiu F.,
- et al.
- Saksena S.,
- Prakash A.,
- Hill M.,
- et al.
- Gerstenfeld E.P.,
- Hill M.R.S.,
- French S.N.,
- et al.
- Coumel P.
- Luck J.C.,
- Engel T.R.