| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department of Anesthesia and Perioperative Care; the
Department of Medicine, Section of Cardiac Electrophysiology, University of California, San Francisco, CA, and the
Department of Cardiovascular Anesthesia, Kaiser Permanente Medical Center, San Francisco, California
Address correspondence to Jacqueline M. Leung, MD, MPH, University of California, San Francisco, Mount Zion Medical Center, Department of Anesthesia and Perioperative Care, 1600 Divisadero Street, Room C-355, San Francisco, CA 94115. Address email to jmleung{at}itsa.ucsf.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
IMPLICATIONS: In addition to clinical factors, such as advanced age and body surface area, we demonstrated that electrophysiologic changes involving an increase in P-wave dispersion postoperatively independently predict atrial fibrillation after coronary artery bypass graft surgery.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
The pathophysiologic basis for the development of AF after coronary artery bypass graft (CABG) surgery is not known despite previous studies that have identified certain clinical predictors of postoperative AF (3,5,7). In addition to clinical factors, an electrophysiologic basis for AF has been investigated. In the aging heart, the sinoatrial (SA) node heart muscle may be replaced by fibrous tissue (8), which may result in changes in cardiac conduction and lead to re-entry of atrial excitation and induction of AF. Furthermore, basic dispersion of refractoriness increases with age, and human electrophysiologic studies have shown that both basic refractoriness and its dispersion are important factors in the occurrence of AF (9).
In addition to preexistent patient conditions, surgical factors that may result in atrial inflammation and/or ischemia may increase the likelihood of postoperative AF. These atrial changes may result in intra-atrial conduction abnormalities detectable in P-wave characteristics on surface electrocardiogram (ECG). P-wave duration and P-wave dispersion on standard ECG are noninvasive markers of intraatrial conduction disturbances, which are believed to be the main electrophysiological cause of AF (10–16); however, no large study has investigated the impact of cardiac surgery on atrial electrophysiology relative to clinical conditions in predicting postoperative AF.
Our principal hypothesis is that interactions between aging-related atrial changes, and acute surgical factors, lead to the development of postoperative AF after cardiac surgery. Acute surgical factors superimposed on aging-related atrial changes result in electrophysiologic changes that can be measured and predictive of the subsequent occurrence of postoperative AF. Accordingly, the purpose of this study was to examine the effect of surgery on pre- and postoperative atrial electrophysiology as measured by P-wave analysis and to determine the potential predictive value of P-wave changes relative to clinical factors on the incidence of postoperative AF in patients undergoing CABG surgery.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Standard monitors for patients undergoing CABG surgery were used. These included 5-lead ECG, radial and pulmonary arterial (PA) catheters, and transesophageal echocardiography. Anesthetic and surgical management of patients was not controlled and was per usual clinical practice.
Cardiopulmonary bypass (CPB) was conducted using a membrane oxygenator with hemodilution and mild systemic hypothermia. Multidose cardioplegia was used for myocardial protection during CPB. Distal coronary anastomoses were usually performed first during continuous aortic cross-clamping, followed by proximal vein grafting during partial aortic occlusion.
A preoperative 12-lead ECG obtained at the time of preoperative evaluation, within 1 wk of the scheduled surgery, and postoperative ECG on either day 1 or 2 after surgery were recorded in all patients at a paper speed of 25 mm/s using a Hewlett Packard M1700A ECG recorder. The frequency response of the system was set at 0.15–150 Hz. All ECG tracings were calibrated to 1 cm = 10 mV (standard). All P-wave measurements were performed in the lead II rhythm strip manually using a standard metric ruler and magnifying lens that was found to be more precise than the standard caliper method for QT intervals (17). All measurable P-waves from lead II rhythm strips were used to determine the P-wave characteristics. The onset of the P-wave was defined as the junction between the isoelectric line and the beginning of P-wave deflection and the offset of the P-wave was defined as the junction between the end of the P-wave deflection and the isoelectric line (18). P-wave duration was measured from the onset to the offset of the P-wave. P-wave amplitude was measured as the absolute value from the isoelectric baseline to positive or negative maximum deflection (19). Biphasic P-wave amplitude was defined as the sum of both the absolute values of positive and negative deflections. P-wave axis was determined by the ECG recorder. PR segment depression was considered significant if any deviation from baseline occurred. PR interval was measured as the time interval from the beginning of the P-wave to the beginning of the QRS complex (19). P-wave dispersion, a measurement of the heterogeneity of atrial depolarizations, was measured as the difference between the duration of the longest and the shortest P-waves in lead II (17). A rhythm strip from one patient demonstrating evidence of P-wave dispersion is shown in Figure 1. All final values were derived from the average of two measurements recorded for each variable. One investigator performed all of the measurements, and a second investigator (a cardiologist) validated the accuracy using a random sample of 20% of the ECGs. Bias analysis (the mean difference between 2 methods) (20) was used to statistically compare the agreement between measurements from the 2 investigators. Both investigators performing the ECG measurements were blinded to the clinical outcomes.
|
All patients were followed daily with continuous ECG telemetry for the occurrence of postoperative AF during the in-hospital period until discharge. AF was defined as AF lasting for a minimum duration of 30 minutes. ECG evidence for pericarditis was determined by ST segment elevation of 1 mm or more in all leads except aVR and V1 (24). Two investigators validated all outcomes by hard copy printout. Postoperative myocardial infarction was defined as new Q waves (
40 ms, 25% R-wave) on postoperative 12-lead ECG and creatine phosphokinase-myocardial band isoenzyme concentration 50 U/L.
Data were analyzed using Stata 5.0 (Stata, College Station, TX). Differences between pre- and postoperative P-wave variables were evaluated using Student’s t-test. The association of postoperative AF and potential clinical predictors was determined by univariate analysis using 
2 test or Student’s t-test. The predictors (covariates) evaluated included patient demographic data such as age, gender, and body surface area (BSA); co-existent medical conditions such as hypertension, history of myocardial infarction, congestive heart failure, or AF; preoperative medication use such as antiarrhythmic drugs and ß blocking drugs; surgical factors such as aortic cross-clamp and bypass times, number of grafts, use of intraoperative defibrillator, adequacy of myocardial protection; and postoperative conditions such as return to the operating room and perioperative myocardial infarction. Those predictors that had significant association with postoperative AF on univariate analysis (P 
0.1) were entered in a stepwise multivariate logistic regression model. After the general model including only the clinical predictors was determined, the incremental value of ECG data was determined by adding ECG P-wave variables to the general model to determine if interactions occurred, and whether the fit of the model was improved with the ECG data using the Pearson and Hosmer-Lemeshow goodness-of fit-tests. Odds ratios (OR) and 95% confidence interval (CI) were reported. P < 0.05 was considered statistically significant. Unless stated otherwise, all data are presented as mean ± SD.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
|
ECG evidence for pericarditis on any postoperative ECG was relatively infrequent, occurring only in eight patients, two of whom developed postoperative AF. The presence of either ST depression or elevation on preoperative ECG, or perioperative myocardial infarction was not associated with postoperative AF. Multivariate logistic regression analysis including only the clinical variables demonstrated that age, BSA, and a history of AF were independent predictors of postoperative AF (Table 2). However, the confidence intervals for BSA and a history of AF are large, indicating the estimates for these predictors are not precise.
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Clinical Predictors
Our finding that 27% of patients developed postoperative AF is similar to previous reports (3,5,7,25,26). Our results also agree with those from previous studies demonstrating that advanced age is an independent predictor of AF. Aging may lead to the loss of myocardial fibers and an increase in fibrosis (27), causing heterogeneity of the structural and electrophysiological properties of the atrial myocardium. This may, in turn, alter the atrial refractory period and conduction velocity leading to an increase in multiple reentrant wavelets as proposed by Moe (28), and experimentally supported in the canine model by Allessie et al. (29).
Our study also found BSA was an independent predictor of postoperative AF, a finding similar to that reported previously (30). However, the CI for the estimate was large, suggesting that the precision of this estimate needs confirmation. Speculations for the association include large BSA being a marker for atrial remodeling with adipose tissue constituting a fertile anatomical substrate for the development of postoperative AF (31). Another hypothesis is that individuals with increased BSA may have larger atria and/or differences in intrathoracic pressure, which may alter the electrophysiological properties of the atrium, making them susceptible to atrial arrhythmias after surgery (32).
Electrophysiological Mechanisms of AF
The electrophysiological mechanism of AF has been under intense investigation for a long time. A single automatic focus firing impulses at a rapid rate, such as that occurring in the pulmonary vein (33) and micro-re-entrant circuits, are possible mechanisms (34). Cellular and electrophysiologic abnormalities necessary for the genesis of AF include fibrosis, partly depolarized cells, conduction abnormalities, shortening of the refractoriness, and increase in dispersion of refractoriness in the atrium (35). Noninvasive ECG markers, such as P-wave duration, a surrogate marker of the atrial conduction velocity, and P-wave dispersion, a surrogate measure of the refractoriness of adjacent atrium, have been most commonly investigated.
Steinberg et al. (11) used signal-averaged ECG and demonstrated that P-wave duration was significantly longer in the patients with postoperative AF; however, the impact of how surgery may have altered this electrophysiologic phenomenon was not investigated, as only preoperative ECG were evaluated. In contrast, Raitt and Ingram (36), using P-wave signal averaging, concluded that P-wave duration was significantly shortened after open-heart surgery. Our study agrees with this finding, as does Tsikouris et al. (37), who also found a decrease in several P-wave features after surgery. However, both of these previous studies evaluated few patients and did not investigate whether postoperative changes in P-wave characteristics were associated with postoperative AF. Increases in postoperative ß-adrenergic tone may account for decreases in P-wave duration, amplitude, and PR interval on immediate postoperative ECG. Cheema et al. (38), demonstrated that ß-adrenergic stimulation with isoproterenol resulted in significant shortening of the signal-averaged P-wave duration, whereas ß-adrenergic blockade resulted in significant prolongation of the P-wave duration. However, P-wave duration measured pre- or postoperatively was not a predictor of postoperative AF in our study, a finding similarly reported from previous studies (32,39).
In contrast, our study shows that an increase in P-wave dispersion postoperatively independently predicts postoperative AF after CABG surgery, a result also seen in other patient populations. For example, Yamada et al. (40) investigated whether the spatial dispersion of signal-averaged P-wave duration was increased in patients with paroxysmal AF, compared with control subjects. These authors reported that the dispersion was significantly greater in the patients with paroxysmal AF than the controls (26.6 ± 9.5 versus 14.8 ± 6.7 ms, P < 0.0001). The mean P-wave dispersion values in this report are similar to those reported in our study. In another study of patients who have suffered an anterior myocardial infarction (41), the authors reported that age >65 years, P-wave duration >125 ms, and P-wave dispersion >25 ms were independently associated with AF. Many of these results are similar to our current findings despite the study populations being different.
P-wave dispersion has been advocated as a novel measurement of the heterogeneity of atrial depolarizations (42). Animal studies have demonstrated that dispersion of refractoriness is one of the essential elements for sustaining atrial arrhythmia (43). The increase in dispersion of atrial refractoriness may have favored a re-entry mechanism leading to AF because fragmentation of the depolarizing wavefront may occur when it faces adjoining atrium with nonuniform refractoriness. However, the electrophysiological mechanism of postoperative AF is likely multifactorial, where abnormal automaticity and atrial conduction delay might be additional electrophysiological substrates (34).
Possible Mechanisms of Postoperative Increase in P-Wave Dispersion
Mechanisms leading to the postoperative increase in P-wave dispersion remains unclear. One proposed etiology for postoperative AF and changes in P-wave characteristics may be fluid overload contributing to atrial stretch (37). In fluid overloaded conditions such as congestive heart failure, P-wave duration has been found to be directly related to left atrial pressure (44). In addition, acute diuresis has been shown to decrease P-wave duration and dispersion in patients with decompensated heart failure (45). Hravnak et al. (30) noted that patients with postoperative AF had a larger positive net fluid balance on the operative day compared with patients who did not develop AF. Inhomogeneous effects of pressure related atrial stretch might result in dispersion of atrial refractoriness and produce local areas of conduction block, thus facilitating reentrant wavelets and the induction of AF. We examined this possibility indirectly in a subset of the same study cohort by estimating left atrial pressure using PA pressure measurements up to the first postoperative day. Overall, substantial PA diastolic pressure increase (>20 mm Hg) was infrequent, occurring in only 14.5% of all intra-and postoperative monitoring times, and it was not associated with postoperative AF.
In addition to loading condition changes, other processes may have resulted in an increase in P-wave dispersion after CABG surgery. In patients undergoing transluminal coronary angioplasty, balloon occlusion of the right coronary artery, known to supply the atrial wall and the atrioventricular and SA nodes, resulted in increases in P-wave duration and dispersion (46). The authors speculated that the increased P-wave duration may have been secondary to ischemia-induced increases in left ventricular end-diastolic pressure and consequent increases in left atrial pressure and volume. Similar changes in P-wave dispersion have been observed in patients with spontaneous anginal episodes (47). It is postulated that in the acutely ischemic myocardium, the rapid accumulation of Ca2+ and H+ leads to a significant decrease in junctional coupling, thereby creating inhomogeneous and discontinuous areas of atrial conduction, leading to increases in P-wave dispersion. Whether the increases in P-wave dispersion shown in our study were a manifestation of atrial ischemia remains to be determined. Indirectly however, our study demonstrated that the myocardial protection as assessed clinically was similar between patients with and without postoperative AF.
Pericarditis has been reported to be one of the factors associated with AF, especially in the setting of acute myocardial infarction (48). We examined this hypothesis in the postoperative CABG setting by reviewing ECG markers for pericarditis such as diffuse ST segment elevation observed in postoperative ECG. Overall, the incidence of ECG evidence for pericarditis was relatively small (8 of 300) and only 2 of those patients developed AF. We additionally examined PR segment depression, which has been associated with pericarditis in a previous study (49), and found that factor was not associated with the development of postoperative AF.
More recently, perioperative magnesium treatment has been shown to reduce P-wave dispersion postoperatively and the occurrence of postoperative AF (50). This result provides a plausible etiology for AF after cardiac surgery but needs validation by further studies.
In addition to an electrophysiologic basis for AF, age-related changes in the atria are additional substrates leading to the development in postoperative AF. Our recent report that left atrial enlargement was associated with postoperative AF, suggests that a preexistent substrate is necessary for the development of postoperative AF (51).
Potential Study Limitations
We measured postoperative P-wave characteristics using only one postoperative ECG recording from a single lead. Serial ECG and multiple ECG leads may have provided more information about the natural history of changes in P-wave characteristics during the postoperative period. However, because most episodes of postoperative AF in our study occurred by postoperative days 2 or 3, evaluating P-wave characteristics in the early postoperative period (postoperative days 1 or 2) likely yields the most useful clinical information in predicting subsequent AF.
The postoperative ECGs in our study were not gathered at the same time of day, and the changes may reflect effects attributable to circadian variation. However, the relative importance of circadian variation on P-wave characteristics is unknown after CABG surgery (52).
Our study only focused on the early postoperative period; as a result, our study may have underestimated the incidence of postoperative AF that may have occurred in the late postoperative period.
Finally, high-resolution ECG or signal averaging was not used in this study and may have increased the accuracy and precision of our measurements for P-wave duration over the manual method (53). However, P-wave signal averaging has other limitations, such as the variability of the PR interval and inclusion of atrial ectopic beats that may render the averaging of P waves less accurate when triggered on the R wave, but the accuracy is also decreased when the onset of the P-wave is used as the trigger (14).
Clinical Implications
Contrary to our original hypothesis, we were unable to readily identify any surgical factor that may be associated with atrial ischemia or injury resulting in postoperative AF. The investigation of a "triggering" factor, however, should be continued with more sensitive measurement methods given the prevalence of AF after cardiothoracic surgery. In contrast, results from the current study as well as those from others suggest that preexistent substrates such as old age, a large atrium, and increased P-wave dispersion postoperatively are necessary for the development of postoperative AF. Our results suggest that the etiology for postoperative AF is likely multifactorial and still incompletely understood. Of the three independent predictors identified, the effect size of age and increases in P-wave dispersion is small, and the predictive value of BSA has a large CI. As a result, direct extrapolation of our results to clinical practice should be cautiously applied. Our results do indicate that a selective therapeutic approach involving clinical risk prediction (advanced age, left atrial enlargement, and increase in postoperative P-wave dispersion) should be further investigated as a possible alternative choice to the nonjudicial use of prophylaxis in all patients undergoing cardiac surgery because the latter approach is likely to be prohibitive in cost and also may yield potential unwanted side effects.
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Kaireviciute, A. Aidietis, and G. Y.H. Lip Atrial fibrillation following cardiac surgery: clinical features and preventative strategies Eur. Heart J., February 2, 2009; 30(4): 410 - 425. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Magnani, M. A. Williamson, P. T. Ellinor, K. M. Monahan, and E. J. Benjamin P Wave Indices: Current Status and Future Directions in Epidemiology, Clinical, and Research Applications Circ Arrhythmia Electrophysiol, February 1, 2009; 2(1): 72 - 79. [Full Text] [PDF]
|
|
|
|
 |
|
 C. Materazzo, P. Piotti, C. Mantovani, R. Miceli, and F. Villani Atrial fibrillation after non-cardiac surgery: P-wave characteristics and Holter monitoring in risk assessment Eur. J. Cardiothorac. Surg., May 1, 2007; 31(5): 812 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Hosokawa, Y. Nakajima, T. Umenai, H. Ueno, S. Taniguchi, T. Matsukawa, and T. Mizobe Predictors of atrial fibrillation after off-pump coronary artery bypass graft surgery Br. J. Anaesth., May 1, 2007; 98(5): 575 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. G. Koch, L. Li, D. R. Van Wagoner, A. I. Duncan, A. M. Gillinov, and E. H. Blackstone Red Cell Transfusion is Associated With an Increased Risk for Postoperative Atrial Fibrillation Ann. Thorac. Surg., November 1, 2006; 82(5): 1747 - 1756. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Leung, W. H. Bellows, and N. B. Schiller Impairment of left atrial function predicts post-operative atrial fibrillation after coronary artery bypass graft surgery Eur. Heart J., October 2, 2004; 25(20): 1836 - 1844. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
